// SPDX-License-Identifier: GPL-2.0-only #include "amd64_edac.h" #include static struct edac_pci_ctl_info *pci_ctl; /* * Set by command line parameter. If BIOS has enabled the ECC, this override is * cleared to prevent re-enabling the hardware by this driver. */ static int ecc_enable_override; module_param(ecc_enable_override, int, 0644); static struct msr __percpu *msrs; static inline u32 get_umc_reg(struct amd64_pvt *pvt, u32 reg) { if (!pvt->flags.zn_regs_v2) return reg; switch (reg) { case UMCCH_ADDR_CFG: return UMCCH_ADDR_CFG_DDR5; case UMCCH_ADDR_MASK_SEC: return UMCCH_ADDR_MASK_SEC_DDR5; case UMCCH_DIMM_CFG: return UMCCH_DIMM_CFG_DDR5; } WARN_ONCE(1, "%s: unknown register 0x%x", __func__, reg); return 0; } /* Per-node stuff */ static struct ecc_settings **ecc_stngs; /* Device for the PCI component */ static struct device *pci_ctl_dev; /* * Valid scrub rates for the K8 hardware memory scrubber. We map the scrubbing * bandwidth to a valid bit pattern. The 'set' operation finds the 'matching- * or higher value'. * *FIXME: Produce a better mapping/linearisation. */ static const struct scrubrate { u32 scrubval; /* bit pattern for scrub rate */ u32 bandwidth; /* bandwidth consumed (bytes/sec) */ } scrubrates[] = { { 0x01, 1600000000UL}, { 0x02, 800000000UL}, { 0x03, 400000000UL}, { 0x04, 200000000UL}, { 0x05, 100000000UL}, { 0x06, 50000000UL}, { 0x07, 25000000UL}, { 0x08, 12284069UL}, { 0x09, 6274509UL}, { 0x0A, 3121951UL}, { 0x0B, 1560975UL}, { 0x0C, 781440UL}, { 0x0D, 390720UL}, { 0x0E, 195300UL}, { 0x0F, 97650UL}, { 0x10, 48854UL}, { 0x11, 24427UL}, { 0x12, 12213UL}, { 0x13, 6101UL}, { 0x14, 3051UL}, { 0x15, 1523UL}, { 0x16, 761UL}, { 0x00, 0UL}, /* scrubbing off */ }; int __amd64_read_pci_cfg_dword(struct pci_dev *pdev, int offset, u32 *val, const char *func) { int err = 0; err = pci_read_config_dword(pdev, offset, val); if (err) amd64_warn("%s: error reading F%dx%03x.\n", func, PCI_FUNC(pdev->devfn), offset); return err; } int __amd64_write_pci_cfg_dword(struct pci_dev *pdev, int offset, u32 val, const char *func) { int err = 0; err = pci_write_config_dword(pdev, offset, val); if (err) amd64_warn("%s: error writing to F%dx%03x.\n", func, PCI_FUNC(pdev->devfn), offset); return err; } /* * Select DCT to which PCI cfg accesses are routed */ static void f15h_select_dct(struct amd64_pvt *pvt, u8 dct) { u32 reg = 0; amd64_read_pci_cfg(pvt->F1, DCT_CFG_SEL, ®); reg &= (pvt->model == 0x30) ? ~3 : ~1; reg |= dct; amd64_write_pci_cfg(pvt->F1, DCT_CFG_SEL, reg); } /* * * Depending on the family, F2 DCT reads need special handling: * * K8: has a single DCT only and no address offsets >= 0x100 * * F10h: each DCT has its own set of regs * DCT0 -> F2x040.. * DCT1 -> F2x140.. * * F16h: has only 1 DCT * * F15h: we select which DCT we access using F1x10C[DctCfgSel] */ static inline int amd64_read_dct_pci_cfg(struct amd64_pvt *pvt, u8 dct, int offset, u32 *val) { switch (pvt->fam) { case 0xf: if (dct || offset >= 0x100) return -EINVAL; break; case 0x10: if (dct) { /* * Note: If ganging is enabled, barring the regs * F2x[1,0]98 and F2x[1,0]9C; reads reads to F2x1xx * return 0. (cf. Section 2.8.1 F10h BKDG) */ if (dct_ganging_enabled(pvt)) return 0; offset += 0x100; } break; case 0x15: /* * F15h: F2x1xx addresses do not map explicitly to DCT1. * We should select which DCT we access using F1x10C[DctCfgSel] */ dct = (dct && pvt->model == 0x30) ? 3 : dct; f15h_select_dct(pvt, dct); break; case 0x16: if (dct) return -EINVAL; break; default: break; } return amd64_read_pci_cfg(pvt->F2, offset, val); } /* * Memory scrubber control interface. For K8, memory scrubbing is handled by * hardware and can involve L2 cache, dcache as well as the main memory. With * F10, this is extended to L3 cache scrubbing on CPU models sporting that * functionality. * * This causes the "units" for the scrubbing speed to vary from 64 byte blocks * (dram) over to cache lines. This is nasty, so we will use bandwidth in * bytes/sec for the setting. * * Currently, we only do dram scrubbing. If the scrubbing is done in software on * other archs, we might not have access to the caches directly. */ /* * 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 __set_scrub_rate(struct amd64_pvt *pvt, u32 new_bw, u32 min_rate) { u32 scrubval; int i; /* * map the configured rate (new_bw) to a value specific to the AMD64 * memory controller and apply to register. Search for the first * bandwidth entry that is greater or equal than the setting requested * and program that. If at last entry, turn off DRAM scrubbing. * * If no suitable bandwidth is found, turn off DRAM scrubbing entirely * by falling back to the last element in scrubrates[]. */ for (i = 0; i < ARRAY_SIZE(scrubrates) - 1; i++) { /* * skip scrub rates which aren't recommended * (see F10 BKDG, F3x58) */ if (scrubrates[i].scrubval < min_rate) continue; if (scrubrates[i].bandwidth <= new_bw) break; } scrubval = scrubrates[i].scrubval; if (pvt->fam == 0x15 && pvt->model == 0x60) { f15h_select_dct(pvt, 0); pci_write_bits32(pvt->F2, F15H_M60H_SCRCTRL, scrubval, 0x001F); f15h_select_dct(pvt, 1); pci_write_bits32(pvt->F2, F15H_M60H_SCRCTRL, scrubval, 0x001F); } else { pci_write_bits32(pvt->F3, SCRCTRL, scrubval, 0x001F); } if (scrubval) return scrubrates[i].bandwidth; return 0; } static int set_scrub_rate(struct mem_ctl_info *mci, u32 bw) { struct amd64_pvt *pvt = mci->pvt_info; u32 min_scrubrate = 0x5; if (pvt->fam == 0xf) min_scrubrate = 0x0; if (pvt->fam == 0x15) { /* Erratum #505 */ if (pvt->model < 0x10) f15h_select_dct(pvt, 0); if (pvt->model == 0x60) min_scrubrate = 0x6; } return __set_scrub_rate(pvt, bw, min_scrubrate); } static int get_scrub_rate(struct mem_ctl_info *mci) { struct amd64_pvt *pvt = mci->pvt_info; int i, retval = -EINVAL; u32 scrubval = 0; if (pvt->fam == 0x15) { /* Erratum #505 */ if (pvt->model < 0x10) f15h_select_dct(pvt, 0); if (pvt->model == 0x60) amd64_read_pci_cfg(pvt->F2, F15H_M60H_SCRCTRL, &scrubval); else amd64_read_pci_cfg(pvt->F3, SCRCTRL, &scrubval); } else { amd64_read_pci_cfg(pvt->F3, SCRCTRL, &scrubval); } scrubval = scrubval & 0x001F; for (i = 0; i < ARRAY_SIZE(scrubrates); i++) { if (scrubrates[i].scrubval == scrubval) { retval = scrubrates[i].bandwidth; break; } } return retval; } /* * returns true if the SysAddr given by sys_addr matches the * DRAM base/limit associated with node_id */ static bool base_limit_match(struct amd64_pvt *pvt, u64 sys_addr, u8 nid) { u64 addr; /* The K8 treats this as a 40-bit value. However, bits 63-40 will be * all ones if the most significant implemented address bit is 1. * Here we discard bits 63-40. See section 3.4.2 of AMD publication * 24592: AMD x86-64 Architecture Programmer's Manual Volume 1 * Application Programming. */ addr = sys_addr & 0x000000ffffffffffull; return ((addr >= get_dram_base(pvt, nid)) && (addr <= get_dram_limit(pvt, nid))); } /* * Attempt to map a SysAddr to a node. On success, return a pointer to the * mem_ctl_info structure for the node that the SysAddr maps to. * * On failure, return NULL. */ static struct mem_ctl_info *find_mc_by_sys_addr(struct mem_ctl_info *mci, u64 sys_addr) { struct amd64_pvt *pvt; u8 node_id; u32 intlv_en, bits; /* * Here we use the DRAM Base (section 3.4.4.1) and DRAM Limit (section * 3.4.4.2) registers to map the SysAddr to a node ID. */ pvt = mci->pvt_info; /* * The value of this field should be the same for all DRAM Base * registers. Therefore we arbitrarily choose to read it from the * register for node 0. */ intlv_en = dram_intlv_en(pvt, 0); if (intlv_en == 0) { for (node_id = 0; node_id < DRAM_RANGES; node_id++) { if (base_limit_match(pvt, sys_addr, node_id)) goto found; } goto err_no_match; } if (unlikely((intlv_en != 0x01) && (intlv_en != 0x03) && (intlv_en != 0x07))) { amd64_warn("DRAM Base[IntlvEn] junk value: 0x%x, BIOS bug?\n", intlv_en); return NULL; } bits = (((u32) sys_addr) >> 12) & intlv_en; for (node_id = 0; ; ) { if ((dram_intlv_sel(pvt, node_id) & intlv_en) == bits) break; /* intlv_sel field matches */ if (++node_id >= DRAM_RANGES) goto err_no_match; } /* sanity test for sys_addr */ 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); return NULL; } found: return edac_mc_find((int)node_id); err_no_match: edac_dbg(2, "sys_addr 0x%lx doesn't match any node\n", (unsigned long)sys_addr); return NULL; } /* * compute the CS base address of the @csrow on the DRAM controller @dct. * For details see F2x[5C:40] in the processor's BKDG */ static void get_cs_base_and_mask(struct amd64_pvt *pvt, int csrow, u8 dct, u64 *base, u64 *mask) { u64 csbase, csmask, base_bits, mask_bits; u8 addr_shift; 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_ULL(31, 21) | GENMASK_ULL(15, 9); mask_bits = GENMASK_ULL(29, 21) | GENMASK_ULL(15, 9); addr_shift = 4; /* * F16h and F15h, models 30h and later need two addr_shift values: * 8 for high and 6 for low (cf. F16h BKDG). */ } else if (pvt->fam == 0x16 || (pvt->fam == 0x15 && pvt->model >= 0x30)) { csbase = pvt->csels[dct].csbases[csrow]; csmask = pvt->csels[dct].csmasks[csrow >> 1]; *base = (csbase & GENMASK_ULL(15, 5)) << 6; *base |= (csbase & GENMASK_ULL(30, 19)) << 8; *mask = ~0ULL; /* poke holes for the csmask */ *mask &= ~((GENMASK_ULL(15, 5) << 6) | (GENMASK_ULL(30, 19) << 8)); *mask |= (csmask & GENMASK_ULL(15, 5)) << 6; *mask |= (csmask & GENMASK_ULL(30, 19)) << 8; return; } else { csbase = pvt->csels[dct].csbases[csrow]; csmask = pvt->csels[dct].csmasks[csrow >> 1]; addr_shift = 8; if (pvt->fam == 0x15) base_bits = mask_bits = GENMASK_ULL(30,19) | GENMASK_ULL(13,5); else base_bits = mask_bits = GENMASK_ULL(28,19) | GENMASK_ULL(13,5); } *base = (csbase & base_bits) << addr_shift; *mask = ~0ULL; /* poke holes for the csmask */ *mask &= ~(mask_bits << addr_shift); /* OR them in */ *mask |= (csmask & mask_bits) << addr_shift; } #define for_each_chip_select(i, dct, pvt) \ for (i = 0; i < pvt->csels[dct].b_cnt; i++) #define chip_select_base(i, dct, pvt) \ pvt->csels[dct].csbases[i] #define for_each_chip_select_mask(i, dct, pvt) \ for (i = 0; i < pvt->csels[dct].m_cnt; i++) #define for_each_umc(i) \ for (i = 0; i < pvt->max_mcs; i++) /* * @input_addr is an InputAddr associated with the node given by mci. Return the * csrow that input_addr maps to, or -1 on failure (no csrow claims input_addr). */ static int input_addr_to_csrow(struct mem_ctl_info *mci, u64 input_addr) { struct amd64_pvt *pvt; int csrow; u64 base, mask; pvt = mci->pvt_info; for_each_chip_select(csrow, 0, pvt) { if (!csrow_enabled(csrow, 0, pvt)) continue; get_cs_base_and_mask(pvt, csrow, 0, &base, &mask); mask = ~mask; if ((input_addr & mask) == (base & mask)) { edac_dbg(2, "InputAddr 0x%lx matches csrow %d (node %d)\n", (unsigned long)input_addr, csrow, pvt->mc_node_id); return csrow; } } edac_dbg(2, "no matching csrow for InputAddr 0x%lx (MC node %d)\n", (unsigned long)input_addr, pvt->mc_node_id); return -1; } /* * Obtain info from the DRAM Hole Address Register (section 3.4.8, pub #26094) * for the node represented by mci. Info is passed back in *hole_base, * *hole_offset, and *hole_size. Function returns 0 if info is valid or 1 if * info is invalid. Info may be invalid for either of the following reasons: * * - The revision of the node is not E or greater. In this case, the DRAM Hole * Address Register does not exist. * * - The DramHoleValid bit is cleared in the DRAM Hole Address Register, * indicating that its contents are not valid. * * The values passed back in *hole_base, *hole_offset, and *hole_size are * complete 32-bit values despite the fact that the bitfields in the DHAR * only represent bits 31-24 of the base and offset values. */ static int get_dram_hole_info(struct mem_ctl_info *mci, u64 *hole_base, u64 *hole_offset, u64 *hole_size) { struct amd64_pvt *pvt = mci->pvt_info; /* only revE and later have the DRAM Hole Address Register */ if (pvt->fam == 0xf && pvt->ext_model < K8_REV_E) { edac_dbg(1, " revision %d for node %d does not support DHAR\n", pvt->ext_model, pvt->mc_node_id); return 1; } /* valid for Fam10h and above */ if (pvt->fam >= 0x10 && !dhar_mem_hoist_valid(pvt)) { edac_dbg(1, " Dram Memory Hoisting is DISABLED on this system\n"); return 1; } if (!dhar_valid(pvt)) { edac_dbg(1, " Dram Memory Hoisting is DISABLED on this node %d\n", pvt->mc_node_id); return 1; } /* This node has Memory Hoisting */ /* +------------------+--------------------+--------------------+----- * | memory | DRAM hole | relocated | * | [0, (x - 1)] | [x, 0xffffffff] | addresses from | * | | | DRAM hole | * | | | [0x100000000, | * | | | (0x100000000+ | * | | | (0xffffffff-x))] | * +------------------+--------------------+--------------------+----- * * Above is a diagram of physical memory showing the DRAM hole and the * relocated addresses from the DRAM hole. As shown, the DRAM hole * starts at address x (the base address) and extends through address * 0xffffffff. The DRAM Hole Address Register (DHAR) relocates the * addresses in the hole so that they start at 0x100000000. */ *hole_base = dhar_base(pvt); *hole_size = (1ULL << 32) - *hole_base; *hole_offset = (pvt->fam > 0xf) ? f10_dhar_offset(pvt) : k8_dhar_offset(pvt); edac_dbg(1, " DHAR info for node %d base 0x%lx offset 0x%lx size 0x%lx\n", pvt->mc_node_id, (unsigned long)*hole_base, (unsigned long)*hole_offset, (unsigned long)*hole_size); return 0; } #ifdef CONFIG_EDAC_DEBUG #define EDAC_DCT_ATTR_SHOW(reg) \ static ssize_t reg##_show(struct device *dev, \ struct device_attribute *mattr, char *data) \ { \ struct mem_ctl_info *mci = to_mci(dev); \ struct amd64_pvt *pvt = mci->pvt_info; \ \ return sprintf(data, "0x%016llx\n", (u64)pvt->reg); \ } EDAC_DCT_ATTR_SHOW(dhar); EDAC_DCT_ATTR_SHOW(dbam0); EDAC_DCT_ATTR_SHOW(top_mem); EDAC_DCT_ATTR_SHOW(top_mem2); static ssize_t dram_hole_show(struct device *dev, struct device_attribute *mattr, char *data) { struct mem_ctl_info *mci = to_mci(dev); u64 hole_base = 0; u64 hole_offset = 0; u64 hole_size = 0; get_dram_hole_info(mci, &hole_base, &hole_offset, &hole_size); return sprintf(data, "%llx %llx %llx\n", hole_base, hole_offset, hole_size); } /* * update NUM_DBG_ATTRS in case you add new members */ static DEVICE_ATTR(dhar, S_IRUGO, dhar_show, NULL); static DEVICE_ATTR(dbam, S_IRUGO, dbam0_show, NULL); static DEVICE_ATTR(topmem, S_IRUGO, top_mem_show, NULL); static DEVICE_ATTR(topmem2, S_IRUGO, top_mem2_show, NULL); static DEVICE_ATTR_RO(dram_hole); static struct attribute *dbg_attrs[] = { &dev_attr_dhar.attr, &dev_attr_dbam.attr, &dev_attr_topmem.attr, &dev_attr_topmem2.attr, &dev_attr_dram_hole.attr, NULL }; static const struct attribute_group dbg_group = { .attrs = dbg_attrs, }; static ssize_t inject_section_show(struct device *dev, struct device_attribute *mattr, char *buf) { struct mem_ctl_info *mci = to_mci(dev); struct amd64_pvt *pvt = mci->pvt_info; return sprintf(buf, "0x%x\n", pvt->injection.section); } /* * store error injection section value which refers to one of 4 16-byte sections * within a 64-byte cacheline * * range: 0..3 */ static ssize_t inject_section_store(struct device *dev, struct device_attribute *mattr, const char *data, size_t count) { struct mem_ctl_info *mci = to_mci(dev); struct amd64_pvt *pvt = mci->pvt_info; unsigned long value; int ret; ret = kstrtoul(data, 10, &value); if (ret < 0) return ret; if (value > 3) { amd64_warn("%s: invalid section 0x%lx\n", __func__, value); return -EINVAL; } pvt->injection.section = (u32) value; return count; } static ssize_t inject_word_show(struct device *dev, struct device_attribute *mattr, char *buf) { struct mem_ctl_info *mci = to_mci(dev); struct amd64_pvt *pvt = mci->pvt_info; return sprintf(buf, "0x%x\n", pvt->injection.word); } /* * store error injection word value which refers to one of 9 16-bit word of the * 16-byte (128-bit + ECC bits) section * * range: 0..8 */ static ssize_t inject_word_store(struct device *dev, struct device_attribute *mattr, const char *data, size_t count) { struct mem_ctl_info *mci = to_mci(dev); struct amd64_pvt *pvt = mci->pvt_info; unsigned long value; int ret; ret = kstrtoul(data, 10, &value); if (ret < 0) return ret; if (value > 8) { amd64_warn("%s: invalid word 0x%lx\n", __func__, value); return -EINVAL; } pvt->injection.word = (u32) value; return count; } static ssize_t inject_ecc_vector_show(struct device *dev, struct device_attribute *mattr, char *buf) { struct mem_ctl_info *mci = to_mci(dev); struct amd64_pvt *pvt = mci->pvt_info; return sprintf(buf, "0x%x\n", pvt->injection.bit_map); } /* * store 16 bit error injection vector which enables injecting errors to the * corresponding bit within the error injection word above. When used during a * DRAM ECC read, it holds the contents of the of the DRAM ECC bits. */ static ssize_t inject_ecc_vector_store(struct device *dev, struct device_attribute *mattr, const char *data, size_t count) { struct mem_ctl_info *mci = to_mci(dev); struct amd64_pvt *pvt = mci->pvt_info; unsigned long value; int ret; ret = kstrtoul(data, 16, &value); if (ret < 0) return ret; if (value & 0xFFFF0000) { amd64_warn("%s: invalid EccVector: 0x%lx\n", __func__, value); return -EINVAL; } pvt->injection.bit_map = (u32) value; return count; } /* * Do a DRAM ECC read. Assemble staged values in the pvt area, format into * fields needed by the injection registers and read the NB Array Data Port. */ static ssize_t inject_read_store(struct device *dev, struct device_attribute *mattr, const char *data, size_t count) { struct mem_ctl_info *mci = to_mci(dev); struct amd64_pvt *pvt = mci->pvt_info; unsigned long value; u32 section, word_bits; int ret; ret = kstrtoul(data, 10, &value); if (ret < 0) return ret; /* Form value to choose 16-byte section of cacheline */ section = F10_NB_ARRAY_DRAM | SET_NB_ARRAY_ADDR(pvt->injection.section); amd64_write_pci_cfg(pvt->F3, F10_NB_ARRAY_ADDR, section); word_bits = SET_NB_DRAM_INJECTION_READ(pvt->injection); /* Issue 'word' and 'bit' along with the READ request */ amd64_write_pci_cfg(pvt->F3, F10_NB_ARRAY_DATA, word_bits); edac_dbg(0, "section=0x%x word_bits=0x%x\n", section, word_bits); return count; } /* * Do a DRAM ECC write. Assemble staged values in the pvt area and format into * fields needed by the injection registers. */ static ssize_t inject_write_store(struct device *dev, struct device_attribute *mattr, const char *data, size_t count) { struct mem_ctl_info *mci = to_mci(dev); struct amd64_pvt *pvt = mci->pvt_info; u32 section, word_bits, tmp; unsigned long value; int ret; ret = kstrtoul(data, 10, &value); if (ret < 0) return ret; /* Form value to choose 16-byte section of cacheline */ section = F10_NB_ARRAY_DRAM | SET_NB_ARRAY_ADDR(pvt->injection.section); amd64_write_pci_cfg(pvt->F3, F10_NB_ARRAY_ADDR, section); word_bits = SET_NB_DRAM_INJECTION_WRITE(pvt->injection); pr_notice_once("Don't forget to decrease MCE polling interval in\n" "/sys/bus/machinecheck/devices/machinecheck/check_interval\n" "so that you can get the error report faster.\n"); on_each_cpu(disable_caches, NULL, 1); /* Issue 'word' and 'bit' along with the READ request */ amd64_write_pci_cfg(pvt->F3, F10_NB_ARRAY_DATA, word_bits); retry: /* wait until injection happens */ amd64_read_pci_cfg(pvt->F3, F10_NB_ARRAY_DATA, &tmp); if (tmp & F10_NB_ARR_ECC_WR_REQ) { cpu_relax(); goto retry; } on_each_cpu(enable_caches, NULL, 1); edac_dbg(0, "section=0x%x word_bits=0x%x\n", section, word_bits); return count; } /* * update NUM_INJ_ATTRS in case you add new members */ static DEVICE_ATTR_RW(inject_section); static DEVICE_ATTR_RW(inject_word); static DEVICE_ATTR_RW(inject_ecc_vector); static DEVICE_ATTR_WO(inject_write); static DEVICE_ATTR_WO(inject_read); static struct attribute *inj_attrs[] = { &dev_attr_inject_section.attr, &dev_attr_inject_word.attr, &dev_attr_inject_ecc_vector.attr, &dev_attr_inject_write.attr, &dev_attr_inject_read.attr, NULL }; static umode_t inj_is_visible(struct kobject *kobj, struct attribute *attr, int idx) { struct device *dev = kobj_to_dev(kobj); struct mem_ctl_info *mci = container_of(dev, struct mem_ctl_info, dev); struct amd64_pvt *pvt = mci->pvt_info; /* Families which have that injection hw */ if (pvt->fam >= 0x10 && pvt->fam <= 0x16) return attr->mode; return 0; } static const struct attribute_group inj_group = { .attrs = inj_attrs, .is_visible = inj_is_visible, }; #endif /* CONFIG_EDAC_DEBUG */ /* * Return the DramAddr that the SysAddr given by @sys_addr maps to. It is * assumed that sys_addr maps to the node given by mci. * * The first part of section 3.4.4 (p. 70) shows how the DRAM Base (section * 3.4.4.1) and DRAM Limit (section 3.4.4.2) registers are used to translate a * SysAddr to a DramAddr. If the DRAM Hole Address Register (DHAR) is enabled, * then it is also involved in translating a SysAddr to a DramAddr. Sections * 3.4.8 and 3.5.8.2 describe the DHAR and how it is used for memory hoisting. * These parts of the documentation are unclear. I interpret them as follows: * * When node n receives a SysAddr, it processes the SysAddr as follows: * * 1. It extracts the DRAMBase and DRAMLimit values from the DRAM Base and DRAM * Limit registers for node n. If the SysAddr is not within the range * specified by the base and limit values, then node n ignores the Sysaddr * (since it does not map to node n). Otherwise continue to step 2 below. * * 2. If the DramHoleValid bit of the DHAR for node n is clear, the DHAR is * disabled so skip to step 3 below. Otherwise see if the SysAddr is within * the range of relocated addresses (starting at 0x100000000) from the DRAM * hole. If not, skip to step 3 below. Else get the value of the * DramHoleOffset field from the DHAR. To obtain the DramAddr, subtract the * offset defined by this value from the SysAddr. * * 3. Obtain the base address for node n from the DRAMBase field of the DRAM * Base register for node n. To obtain the DramAddr, subtract the base * address from the SysAddr, as shown near the start of section 3.4.4 (p.70). */ static u64 sys_addr_to_dram_addr(struct mem_ctl_info *mci, u64 sys_addr) { struct amd64_pvt *pvt = mci->pvt_info; u64 dram_base, hole_base, hole_offset, hole_size, dram_addr; int ret; dram_base = get_dram_base(pvt, pvt->mc_node_id); ret = get_dram_hole_info(mci, &hole_base, &hole_offset, &hole_size); if (!ret) { if ((sys_addr >= (1ULL << 32)) && (sys_addr < ((1ULL << 32) + hole_size))) { /* use DHAR to translate SysAddr to DramAddr */ dram_addr = sys_addr - hole_offset; edac_dbg(2, "using DHAR to translate SysAddr 0x%lx to DramAddr 0x%lx\n", (unsigned long)sys_addr, (unsigned long)dram_addr); return dram_addr; } } /* * Translate the SysAddr to a DramAddr as shown near the start of * section 3.4.4 (p. 70). Although sys_addr is a 64-bit value, the k8 * only deals with 40-bit values. Therefore we discard bits 63-40 of * sys_addr below. If bit 39 of sys_addr is 1 then the bits we * discard are all 1s. Otherwise the bits we discard are all 0s. See * section 3.4.2 of AMD publication 24592: AMD x86-64 Architecture * Programmer's Manual Volume 1 Application Programming. */ 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); return dram_addr; } /* * @intlv_en is the value of the IntlvEn field from a DRAM Base register * (section 3.4.4.1). Return the number of bits from a SysAddr that are used * for node interleaving. */ static int num_node_interleave_bits(unsigned intlv_en) { static const int intlv_shift_table[] = { 0, 1, 0, 2, 0, 0, 0, 3 }; int n; BUG_ON(intlv_en > 7); n = intlv_shift_table[intlv_en]; return n; } /* Translate the DramAddr given by @dram_addr to an InputAddr. */ static u64 dram_addr_to_input_addr(struct mem_ctl_info *mci, u64 dram_addr) { struct amd64_pvt *pvt; int intlv_shift; u64 input_addr; pvt = mci->pvt_info; /* * See the start of section 3.4.4 (p. 70, BKDG #26094, K8, revA-E) * 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_ULL(35, 12)) + (dram_addr & 0xfff); edac_dbg(2, " Intlv Shift=%d DramAddr=0x%lx maps to InputAddr=0x%lx\n", intlv_shift, (unsigned long)dram_addr, (unsigned long)input_addr); return input_addr; } /* * Translate the SysAddr represented by @sys_addr to an InputAddr. It is * assumed that @sys_addr maps to the node given by mci. */ static u64 sys_addr_to_input_addr(struct mem_ctl_info *mci, u64 sys_addr) { u64 input_addr; input_addr = dram_addr_to_input_addr(mci, sys_addr_to_dram_addr(mci, sys_addr)); edac_dbg(2, "SysAddr 0x%lx translates to InputAddr 0x%lx\n", (unsigned long)sys_addr, (unsigned long)input_addr); return input_addr; } /* Map the Error address to a PAGE and PAGE OFFSET. */ static inline void error_address_to_page_and_offset(u64 error_address, struct err_info *err) { err->page = (u32) (error_address >> PAGE_SHIFT); err->offset = ((u32) error_address) & ~PAGE_MASK; } /* * @sys_addr is an error address (a SysAddr) extracted from the MCA NB Address * Low (section 3.6.4.5) and MCA NB Address High (section 3.6.4.6) registers * of a node that detected an ECC memory error. mci represents the node that * the error address maps to (possibly different from the node that detected * the error). Return the number of the csrow that sys_addr maps to, or -1 on * error. */ static int sys_addr_to_csrow(struct mem_ctl_info *mci, u64 sys_addr) { int csrow; csrow = input_addr_to_csrow(mci, sys_addr_to_input_addr(mci, sys_addr)); if (csrow == -1) amd64_mc_err(mci, "Failed to translate InputAddr to csrow for " "address 0x%lx\n", (unsigned long)sys_addr); return csrow; } /* * See AMD PPR DF::LclNodeTypeMap * * This register gives information for nodes of the same type within a system. * * Reading this register from a GPU node will tell how many GPU nodes are in the * system and what the lowest AMD Node ID value is for the GPU nodes. Use this * info to fixup the Linux logical "Node ID" value set in the AMD NB code and EDAC. */ static struct local_node_map { u16 node_count; u16 base_node_id; } gpu_node_map; #define PCI_DEVICE_ID_AMD_MI200_DF_F1 0x14d1 #define REG_LOCAL_NODE_TYPE_MAP 0x144 /* Local Node Type Map (LNTM) fields */ #define LNTM_NODE_COUNT GENMASK(27, 16) #define LNTM_BASE_NODE_ID GENMASK(11, 0) static int gpu_get_node_map(void) { struct pci_dev *pdev; int ret; u32 tmp; /* * Node ID 0 is reserved for CPUs. * Therefore, a non-zero Node ID means we've already cached the values. */ if (gpu_node_map.base_node_id) return 0; pdev = pci_get_device(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_MI200_DF_F1, NULL); if (!pdev) { ret = -ENODEV; goto out; } ret = pci_read_config_dword(pdev, REG_LOCAL_NODE_TYPE_MAP, &tmp); if (ret) goto out; gpu_node_map.node_count = FIELD_GET(LNTM_NODE_COUNT, tmp); gpu_node_map.base_node_id = FIELD_GET(LNTM_BASE_NODE_ID, tmp); out: pci_dev_put(pdev); return ret; } static int fixup_node_id(int node_id, struct mce *m) { /* MCA_IPID[InstanceIdHi] give the AMD Node ID for the bank. */ u8 nid = (m->ipid >> 44) & 0xF; if (smca_get_bank_type(m->extcpu, m->bank) != SMCA_UMC_V2) return node_id; /* Nodes below the GPU base node are CPU nodes and don't need a fixup. */ if (nid < gpu_node_map.base_node_id) return node_id; /* Convert the hardware-provided AMD Node ID to a Linux logical one. */ return nid - gpu_node_map.base_node_id + 1; } /* Protect the PCI config register pairs used for DF indirect access. */ static DEFINE_MUTEX(df_indirect_mutex); /* * Data Fabric Indirect Access uses FICAA/FICAD. * * Fabric Indirect Configuration Access Address (FICAA): Constructed based * on the device's Instance Id and the PCI function and register offset of * the desired register. * * Fabric Indirect Configuration Access Data (FICAD): There are FICAD LO * and FICAD HI registers but so far we only need the LO register. * * Use Instance Id 0xFF to indicate a broadcast read. */ #define DF_BROADCAST 0xFF static int __df_indirect_read(u16 node, u8 func, u16 reg, u8 instance_id, u32 *lo) { struct pci_dev *F4; u32 ficaa; int err = -ENODEV; if (node >= amd_nb_num()) goto out; F4 = node_to_amd_nb(node)->link; if (!F4) goto out; ficaa = (instance_id == DF_BROADCAST) ? 0 : 1; ficaa |= reg & 0x3FC; ficaa |= (func & 0x7) << 11; ficaa |= instance_id << 16; mutex_lock(&df_indirect_mutex); err = pci_write_config_dword(F4, 0x5C, ficaa); if (err) { pr_warn("Error writing DF Indirect FICAA, FICAA=0x%x\n", ficaa); goto out_unlock; } err = pci_read_config_dword(F4, 0x98, lo); if (err) pr_warn("Error reading DF Indirect FICAD LO, FICAA=0x%x.\n", ficaa); out_unlock: mutex_unlock(&df_indirect_mutex); out: return err; } static int df_indirect_read_instance(u16 node, u8 func, u16 reg, u8 instance_id, u32 *lo) { return __df_indirect_read(node, func, reg, instance_id, lo); } static int df_indirect_read_broadcast(u16 node, u8 func, u16 reg, u32 *lo) { return __df_indirect_read(node, func, reg, DF_BROADCAST, lo); } struct addr_ctx { u64 ret_addr; u32 tmp; u16 nid; u8 inst_id; }; static int umc_normaddr_to_sysaddr(u64 norm_addr, u16 nid, u8 umc, u64 *sys_addr) { u64 dram_base_addr, dram_limit_addr, dram_hole_base; u8 die_id_shift, die_id_mask, socket_id_shift, socket_id_mask; u8 intlv_num_dies, intlv_num_chan, intlv_num_sockets; u8 intlv_addr_sel, intlv_addr_bit; u8 num_intlv_bits, hashed_bit; u8 lgcy_mmio_hole_en, base = 0; u8 cs_mask, cs_id = 0; bool hash_enabled = false; struct addr_ctx ctx; memset(&ctx, 0, sizeof(ctx)); /* Start from the normalized address */ ctx.ret_addr = norm_addr; ctx.nid = nid; ctx.inst_id = umc; /* Read D18F0x1B4 (DramOffset), check if base 1 is used. */ if (df_indirect_read_instance(nid, 0, 0x1B4, umc, &ctx.tmp)) goto out_err; /* Remove HiAddrOffset from normalized address, if enabled: */ if (ctx.tmp & BIT(0)) { u64 hi_addr_offset = (ctx.tmp & GENMASK_ULL(31, 20)) << 8; if (norm_addr >= hi_addr_offset) { ctx.ret_addr -= hi_addr_offset; base = 1; } } /* Read D18F0x110 (DramBaseAddress). */ if (df_indirect_read_instance(nid, 0, 0x110 + (8 * base), umc, &ctx.tmp)) goto out_err; /* Check if address range is valid. */ if (!(ctx.tmp & BIT(0))) { pr_err("%s: Invalid DramBaseAddress range: 0x%x.\n", __func__, ctx.tmp); goto out_err; } lgcy_mmio_hole_en = ctx.tmp & BIT(1); intlv_num_chan = (ctx.tmp >> 4) & 0xF; intlv_addr_sel = (ctx.tmp >> 8) & 0x7; dram_base_addr = (ctx.tmp & GENMASK_ULL(31, 12)) << 16; /* {0, 1, 2, 3} map to address bits {8, 9, 10, 11} respectively */ if (intlv_addr_sel > 3) { pr_err("%s: Invalid interleave address select %d.\n", __func__, intlv_addr_sel); goto out_err; } /* Read D18F0x114 (DramLimitAddress). */ if (df_indirect_read_instance(nid, 0, 0x114 + (8 * base), umc, &ctx.tmp)) goto out_err; intlv_num_sockets = (ctx.tmp >> 8) & 0x1; intlv_num_dies = (ctx.tmp >> 10) & 0x3; dram_limit_addr = ((ctx.tmp & GENMASK_ULL(31, 12)) << 16) | GENMASK_ULL(27, 0); intlv_addr_bit = intlv_addr_sel + 8; /* Re-use intlv_num_chan by setting it equal to log2(#channels) */ switch (intlv_num_chan) { case 0: intlv_num_chan = 0; break; case 1: intlv_num_chan = 1; break; case 3: intlv_num_chan = 2; break; case 5: intlv_num_chan = 3; break; case 7: intlv_num_chan = 4; break; case 8: intlv_num_chan = 1; hash_enabled = true; break; default: pr_err("%s: Invalid number of interleaved channels %d.\n", __func__, intlv_num_chan); goto out_err; } num_intlv_bits = intlv_num_chan; if (intlv_num_dies > 2) { pr_err("%s: Invalid number of interleaved nodes/dies %d.\n", __func__, intlv_num_dies); goto out_err; } num_intlv_bits += intlv_num_dies; /* Add a bit if sockets are interleaved. */ num_intlv_bits += intlv_num_sockets; /* Assert num_intlv_bits <= 4 */ if (num_intlv_bits > 4) { pr_err("%s: Invalid interleave bits %d.\n", __func__, num_intlv_bits); goto out_err; } if (num_intlv_bits > 0) { u64 temp_addr_x, temp_addr_i, temp_addr_y; u8 die_id_bit, sock_id_bit, cs_fabric_id; /* * Read FabricBlockInstanceInformation3_CS[BlockFabricID]. * This is the fabric id for this coherent slave. Use * umc/channel# as instance id of the coherent slave * for FICAA. */ if (df_indirect_read_instance(nid, 0, 0x50, umc, &ctx.tmp)) goto out_err; cs_fabric_id = (ctx.tmp >> 8) & 0xFF; die_id_bit = 0; /* If interleaved over more than 1 channel: */ if (intlv_num_chan) { die_id_bit = intlv_num_chan; cs_mask = (1 << die_id_bit) - 1; cs_id = cs_fabric_id & cs_mask; } sock_id_bit = die_id_bit; /* Read D18F1x208 (SystemFabricIdMask). */ if (intlv_num_dies || intlv_num_sockets) if (df_indirect_read_broadcast(nid, 1, 0x208, &ctx.tmp)) goto out_err; /* If interleaved over more than 1 die. */ if (intlv_num_dies) { sock_id_bit = die_id_bit + intlv_num_dies; die_id_shift = (ctx.tmp >> 24) & 0xF; die_id_mask = (ctx.tmp >> 8) & 0xFF; cs_id |= ((cs_fabric_id & die_id_mask) >> die_id_shift) << die_id_bit; } /* If interleaved over more than 1 socket. */ if (intlv_num_sockets) { socket_id_shift = (ctx.tmp >> 28) & 0xF; socket_id_mask = (ctx.tmp >> 16) & 0xFF; cs_id |= ((cs_fabric_id & socket_id_mask) >> socket_id_shift) << sock_id_bit; } /* * The pre-interleaved address consists of XXXXXXIIIYYYYY * where III is the ID for this CS, and XXXXXXYYYYY are the * address bits from the post-interleaved address. * "num_intlv_bits" has been calculated to tell us how many "I" * bits there are. "intlv_addr_bit" tells us how many "Y" bits * there are (where "I" starts). */ temp_addr_y = ctx.ret_addr & GENMASK_ULL(intlv_addr_bit - 1, 0); temp_addr_i = (cs_id << intlv_addr_bit); temp_addr_x = (ctx.ret_addr & GENMASK_ULL(63, intlv_addr_bit)) << num_intlv_bits; ctx.ret_addr = temp_addr_x | temp_addr_i | temp_addr_y; } /* Add dram base address */ ctx.ret_addr += dram_base_addr; /* If legacy MMIO hole enabled */ if (lgcy_mmio_hole_en) { if (df_indirect_read_broadcast(nid, 0, 0x104, &ctx.tmp)) goto out_err; dram_hole_base = ctx.tmp & GENMASK(31, 24); if (ctx.ret_addr >= dram_hole_base) ctx.ret_addr += (BIT_ULL(32) - dram_hole_base); } if (hash_enabled) { /* Save some parentheses and grab ls-bit at the end. */ hashed_bit = (ctx.ret_addr >> 12) ^ (ctx.ret_addr >> 18) ^ (ctx.ret_addr >> 21) ^ (ctx.ret_addr >> 30) ^ cs_id; hashed_bit &= BIT(0); if (hashed_bit != ((ctx.ret_addr >> intlv_addr_bit) & BIT(0))) ctx.ret_addr ^= BIT(intlv_addr_bit); } /* Is calculated system address is above DRAM limit address? */ if (ctx.ret_addr > dram_limit_addr) goto out_err; *sys_addr = ctx.ret_addr; return 0; out_err: return -EINVAL; } 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 dct_determine_edac_cap(struct amd64_pvt *pvt) { unsigned long edac_cap = EDAC_FLAG_NONE; u8 bit; bit = (pvt->fam > 0xf || pvt->ext_model >= K8_REV_F) ? 19 : 17; if (pvt->dclr0 & BIT(bit)) edac_cap = EDAC_FLAG_SECDED; return edac_cap; } static unsigned long umc_determine_edac_cap(struct amd64_pvt *pvt) { u8 i, umc_en_mask = 0, dimm_ecc_en_mask = 0; unsigned long edac_cap = EDAC_FLAG_NONE; for_each_umc(i) { if (!(pvt->umc[i].sdp_ctrl & UMC_SDP_INIT)) continue; umc_en_mask |= BIT(i); /* UMC Configuration bit 12 (DimmEccEn) */ if (pvt->umc[i].umc_cfg & BIT(12)) dimm_ecc_en_mask |= BIT(i); } if (umc_en_mask == dimm_ecc_en_mask) edac_cap = EDAC_FLAG_SECDED; return edac_cap; } /* * debug routine to display the memory sizes of all logical DIMMs and its * CSROWs */ static void dct_debug_display_dimm_sizes(struct amd64_pvt *pvt, u8 ctrl) { u32 *dcsb = ctrl ? pvt->csels[1].csbases : pvt->csels[0].csbases; u32 dbam = ctrl ? pvt->dbam1 : pvt->dbam0; int dimm, size0, size1; if (pvt->fam == 0xf) { /* K8 families < revF not supported yet */ if (pvt->ext_model < K8_REV_F) return; WARN_ON(ctrl != 0); } if (pvt->fam == 0x10) { dbam = (ctrl && !dct_ganging_enabled(pvt)) ? pvt->dbam1 : pvt->dbam0; dcsb = (ctrl && !dct_ganging_enabled(pvt)) ? pvt->csels[1].csbases : pvt->csels[0].csbases; } else if (ctrl) { dbam = pvt->dbam0; dcsb = pvt->csels[1].csbases; } edac_dbg(1, "F2x%d80 (DRAM Bank Address Mapping): 0x%08x\n", ctrl, dbam); edac_printk(KERN_DEBUG, EDAC_MC, "DCT%d chip selects:\n", ctrl); /* Dump memory sizes for DIMM and its CSROWs */ for (dimm = 0; dimm < 4; dimm++) { size0 = 0; if (dcsb[dimm * 2] & DCSB_CS_ENABLE) /* * For F15m60h, we need multiplier for LRDIMM cs_size * calculation. We pass dimm value to the dbam_to_cs * mapper so we can find the multiplier from the * corresponding DCSM. */ size0 = pvt->ops->dbam_to_cs(pvt, ctrl, DBAM_DIMM(dimm, dbam), dimm); size1 = 0; if (dcsb[dimm * 2 + 1] & DCSB_CS_ENABLE) size1 = pvt->ops->dbam_to_cs(pvt, ctrl, DBAM_DIMM(dimm, dbam), dimm); amd64_info(EDAC_MC ": %d: %5dMB %d: %5dMB\n", dimm * 2, size0, dimm * 2 + 1, size1); } } 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); if (pvt->dram_type == MEM_LRDDR3) { u32 dcsm = pvt->csels[chan].csmasks[0]; /* * It's assumed all LRDIMMs in a DCT are going to be of * same 'type' until proven otherwise. So, use a cs * value of '0' here to get dcsm value. */ edac_dbg(1, " LRDIMM %dx rank multiply\n", (dcsm & 0x3)); } edac_dbg(1, "All DIMMs support ECC:%s\n", (dclr & BIT(19)) ? "yes" : "no"); edac_dbg(1, " PAR/ERR parity: %s\n", (dclr & BIT(8)) ? "enabled" : "disabled"); if (pvt->fam == 0x10) edac_dbg(1, " DCT 128bit mode width: %s\n", (dclr & BIT(11)) ? "128b" : "64b"); edac_dbg(1, " x4 logical DIMMs present: L0: %s L1: %s L2: %s L3: %s\n", (dclr & BIT(12)) ? "yes" : "no", (dclr & BIT(13)) ? "yes" : "no", (dclr & BIT(14)) ? "yes" : "no", (dclr & BIT(15)) ? "yes" : "no"); } #define CS_EVEN_PRIMARY BIT(0) #define CS_ODD_PRIMARY BIT(1) #define CS_EVEN_SECONDARY BIT(2) #define CS_ODD_SECONDARY BIT(3) #define CS_3R_INTERLEAVE BIT(4) #define CS_EVEN (CS_EVEN_PRIMARY | CS_EVEN_SECONDARY) #define CS_ODD (CS_ODD_PRIMARY | CS_ODD_SECONDARY) static int umc_get_cs_mode(int dimm, u8 ctrl, struct amd64_pvt *pvt) { u8 base, count = 0; int cs_mode = 0; if (csrow_enabled(2 * dimm, ctrl, pvt)) cs_mode |= CS_EVEN_PRIMARY; if (csrow_enabled(2 * dimm + 1, ctrl, pvt)) cs_mode |= CS_ODD_PRIMARY; /* Asymmetric dual-rank DIMM support. */ if (csrow_sec_enabled(2 * dimm + 1, ctrl, pvt)) cs_mode |= CS_ODD_SECONDARY; /* * 3 Rank inteleaving support. * There should be only three bases enabled and their two masks should * be equal. */ for_each_chip_select(base, ctrl, pvt) count += csrow_enabled(base, ctrl, pvt); if (count == 3 && pvt->csels[ctrl].csmasks[0] == pvt->csels[ctrl].csmasks[1]) { edac_dbg(1, "3R interleaving in use.\n"); cs_mode |= CS_3R_INTERLEAVE; } return cs_mode; } static int __addr_mask_to_cs_size(u32 addr_mask_orig, unsigned int cs_mode, int csrow_nr, int dimm) { u32 msb, weight, num_zero_bits; u32 addr_mask_deinterleaved; int size = 0; /* * The number of zero bits in the mask is equal to the number of bits * in a full mask minus the number of bits in the current mask. * * The MSB is the number of bits in the full mask because BIT[0] is * always 0. * * In the special 3 Rank interleaving case, a single bit is flipped * without swapping with the most significant bit. This can be handled * by keeping the MSB where it is and ignoring the single zero bit. */ msb = fls(addr_mask_orig) - 1; weight = hweight_long(addr_mask_orig); num_zero_bits = msb - weight - !!(cs_mode & CS_3R_INTERLEAVE); /* Take the number of zero bits off from the top of the mask. */ addr_mask_deinterleaved = GENMASK_ULL(msb - num_zero_bits, 1); edac_dbg(1, "CS%d DIMM%d AddrMasks:\n", csrow_nr, dimm); edac_dbg(1, " Original AddrMask: 0x%x\n", addr_mask_orig); edac_dbg(1, " Deinterleaved AddrMask: 0x%x\n", addr_mask_deinterleaved); /* Register [31:1] = Address [39:9]. Size is in kBs here. */ size = (addr_mask_deinterleaved >> 2) + 1; /* Return size in MBs. */ return size >> 10; } static int umc_addr_mask_to_cs_size(struct amd64_pvt *pvt, u8 umc, unsigned int cs_mode, int csrow_nr) { int cs_mask_nr = csrow_nr; u32 addr_mask_orig; int dimm, size = 0; /* No Chip Selects are enabled. */ if (!cs_mode) return size; /* Requested size of an even CS but none are enabled. */ if (!(cs_mode & CS_EVEN) && !(csrow_nr & 1)) return size; /* Requested size of an odd CS but none are enabled. */ if (!(cs_mode & CS_ODD) && (csrow_nr & 1)) return size; /* * Family 17h introduced systems with one mask per DIMM, * and two Chip Selects per DIMM. * * CS0 and CS1 -> MASK0 / DIMM0 * CS2 and CS3 -> MASK1 / DIMM1 * * Family 19h Model 10h introduced systems with one mask per Chip Select, * and two Chip Selects per DIMM. * * CS0 -> MASK0 -> DIMM0 * CS1 -> MASK1 -> DIMM0 * CS2 -> MASK2 -> DIMM1 * CS3 -> MASK3 -> DIMM1 * * Keep the mask number equal to the Chip Select number for newer systems, * and shift the mask number for older systems. */ dimm = csrow_nr >> 1; if (!pvt->flags.zn_regs_v2) cs_mask_nr >>= 1; /* Asymmetric dual-rank DIMM support. */ if ((csrow_nr & 1) && (cs_mode & CS_ODD_SECONDARY)) addr_mask_orig = pvt->csels[umc].csmasks_sec[cs_mask_nr]; else addr_mask_orig = pvt->csels[umc].csmasks[cs_mask_nr]; return __addr_mask_to_cs_size(addr_mask_orig, cs_mode, csrow_nr, dimm); } static void umc_debug_display_dimm_sizes(struct amd64_pvt *pvt, u8 ctrl) { int dimm, size0, size1, cs0, cs1, cs_mode; edac_printk(KERN_DEBUG, EDAC_MC, "UMC%d chip selects:\n", ctrl); for (dimm = 0; dimm < 2; dimm++) { cs0 = dimm * 2; cs1 = dimm * 2 + 1; cs_mode = umc_get_cs_mode(dimm, ctrl, pvt); size0 = umc_addr_mask_to_cs_size(pvt, ctrl, cs_mode, cs0); size1 = umc_addr_mask_to_cs_size(pvt, ctrl, cs_mode, cs1); amd64_info(EDAC_MC ": %d: %5dMB %d: %5dMB\n", cs0, size0, cs1, size1); } } static void umc_dump_misc_regs(struct amd64_pvt *pvt) { struct amd64_umc *umc; u32 i, tmp, umc_base; for_each_umc(i) { umc_base = get_umc_base(i); umc = &pvt->umc[i]; edac_dbg(1, "UMC%d DIMM cfg: 0x%x\n", i, umc->dimm_cfg); edac_dbg(1, "UMC%d UMC cfg: 0x%x\n", i, umc->umc_cfg); edac_dbg(1, "UMC%d SDP ctrl: 0x%x\n", i, umc->sdp_ctrl); edac_dbg(1, "UMC%d ECC ctrl: 0x%x\n", i, umc->ecc_ctrl); amd_smn_read(pvt->mc_node_id, umc_base + UMCCH_ECC_BAD_SYMBOL, &tmp); edac_dbg(1, "UMC%d ECC bad symbol: 0x%x\n", i, tmp); amd_smn_read(pvt->mc_node_id, umc_base + UMCCH_UMC_CAP, &tmp); edac_dbg(1, "UMC%d UMC cap: 0x%x\n", i, tmp); edac_dbg(1, "UMC%d UMC cap high: 0x%x\n", i, umc->umc_cap_hi); edac_dbg(1, "UMC%d ECC capable: %s, ChipKill ECC capable: %s\n", i, (umc->umc_cap_hi & BIT(30)) ? "yes" : "no", (umc->umc_cap_hi & BIT(31)) ? "yes" : "no"); edac_dbg(1, "UMC%d All DIMMs support ECC: %s\n", i, (umc->umc_cfg & BIT(12)) ? "yes" : "no"); edac_dbg(1, "UMC%d x4 DIMMs present: %s\n", i, (umc->dimm_cfg & BIT(6)) ? "yes" : "no"); edac_dbg(1, "UMC%d x16 DIMMs present: %s\n", i, (umc->dimm_cfg & BIT(7)) ? "yes" : "no"); if (umc->dram_type == MEM_LRDDR4 || umc->dram_type == MEM_LRDDR5) { amd_smn_read(pvt->mc_node_id, umc_base + get_umc_reg(pvt, UMCCH_ADDR_CFG), &tmp); edac_dbg(1, "UMC%d LRDIMM %dx rank multiply\n", i, 1 << ((tmp >> 4) & 0x3)); } umc_debug_display_dimm_sizes(pvt, i); } } static void dct_dump_misc_regs(struct amd64_pvt *pvt) { edac_dbg(1, "F3xE8 (NB Cap): 0x%08x\n", pvt->nbcap); edac_dbg(1, " NB two channel DRAM capable: %s\n", (pvt->nbcap & NBCAP_DCT_DUAL) ? "yes" : "no"); edac_dbg(1, " ECC capable: %s, ChipKill ECC capable: %s\n", (pvt->nbcap & NBCAP_SECDED) ? "yes" : "no", (pvt->nbcap & NBCAP_CHIPKILL) ? "yes" : "no"); debug_dump_dramcfg_low(pvt, pvt->dclr0, 0); edac_dbg(1, "F3xB0 (Online Spare): 0x%08x\n", pvt->online_spare); edac_dbg(1, "F1xF0 (DRAM Hole Address): 0x%08x, base: 0x%08x, offset: 0x%08x\n", pvt->dhar, dhar_base(pvt), (pvt->fam == 0xf) ? k8_dhar_offset(pvt) : f10_dhar_offset(pvt)); dct_debug_display_dimm_sizes(pvt, 0); /* everything below this point is Fam10h and above */ if (pvt->fam == 0xf) return; dct_debug_display_dimm_sizes(pvt, 1); /* Only if NOT ganged does dclr1 have valid info */ if (!dct_ganging_enabled(pvt)) debug_dump_dramcfg_low(pvt, pvt->dclr1, 1); edac_dbg(1, " DramHoleValid: %s\n", dhar_valid(pvt) ? "yes" : "no"); amd64_info("using x%u syndromes.\n", pvt->ecc_sym_sz); } /* * See BKDG, F2x[1,0][5C:40], F2[1,0][6C:60] */ static void dct_prep_chip_selects(struct amd64_pvt *pvt) { if (pvt->fam == 0xf && pvt->ext_model < K8_REV_F) { pvt->csels[0].b_cnt = pvt->csels[1].b_cnt = 8; pvt->csels[0].m_cnt = pvt->csels[1].m_cnt = 8; } else if (pvt->fam == 0x15 && pvt->model == 0x30) { pvt->csels[0].b_cnt = pvt->csels[1].b_cnt = 4; pvt->csels[0].m_cnt = pvt->csels[1].m_cnt = 2; } else { pvt->csels[0].b_cnt = pvt->csels[1].b_cnt = 8; pvt->csels[0].m_cnt = pvt->csels[1].m_cnt = 4; } } static void umc_prep_chip_selects(struct amd64_pvt *pvt) { int umc; for_each_umc(umc) { pvt->csels[umc].b_cnt = 4; pvt->csels[umc].m_cnt = pvt->flags.zn_regs_v2 ? 4 : 2; } } static void umc_read_base_mask(struct amd64_pvt *pvt) { u32 umc_base_reg, umc_base_reg_sec; u32 umc_mask_reg, umc_mask_reg_sec; u32 base_reg, base_reg_sec; u32 mask_reg, mask_reg_sec; u32 *base, *base_sec; u32 *mask, *mask_sec; int cs, umc; for_each_umc(umc) { umc_base_reg = get_umc_base(umc) + UMCCH_BASE_ADDR; umc_base_reg_sec = get_umc_base(umc) + UMCCH_BASE_ADDR_SEC; for_each_chip_select(cs, umc, pvt) { base = &pvt->csels[umc].csbases[cs]; base_sec = &pvt->csels[umc].csbases_sec[cs]; base_reg = umc_base_reg + (cs * 4); base_reg_sec = umc_base_reg_sec + (cs * 4); if (!amd_smn_read(pvt->mc_node_id, base_reg, base)) edac_dbg(0, " DCSB%d[%d]=0x%08x reg: 0x%x\n", umc, cs, *base, base_reg); if (!amd_smn_read(pvt->mc_node_id, base_reg_sec, base_sec)) edac_dbg(0, " DCSB_SEC%d[%d]=0x%08x reg: 0x%x\n", umc, cs, *base_sec, base_reg_sec); } umc_mask_reg = get_umc_base(umc) + UMCCH_ADDR_MASK; umc_mask_reg_sec = get_umc_base(umc) + get_umc_reg(pvt, UMCCH_ADDR_MASK_SEC); for_each_chip_select_mask(cs, umc, pvt) { mask = &pvt->csels[umc].csmasks[cs]; mask_sec = &pvt->csels[umc].csmasks_sec[cs]; mask_reg = umc_mask_reg + (cs * 4); mask_reg_sec = umc_mask_reg_sec + (cs * 4); if (!amd_smn_read(pvt->mc_node_id, mask_reg, mask)) edac_dbg(0, " DCSM%d[%d]=0x%08x reg: 0x%x\n", umc, cs, *mask, mask_reg); if (!amd_smn_read(pvt->mc_node_id, mask_reg_sec, mask_sec)) edac_dbg(0, " DCSM_SEC%d[%d]=0x%08x reg: 0x%x\n", umc, cs, *mask_sec, mask_reg_sec); } } } /* * Function 2 Offset F10_DCSB0; read in the DCS Base and DCS Mask registers */ static void dct_read_base_mask(struct amd64_pvt *pvt) { int cs; for_each_chip_select(cs, 0, pvt) { int reg0 = DCSB0 + (cs * 4); int reg1 = DCSB1 + (cs * 4); u32 *base0 = &pvt->csels[0].csbases[cs]; u32 *base1 = &pvt->csels[1].csbases[cs]; if (!amd64_read_dct_pci_cfg(pvt, 0, reg0, base0)) edac_dbg(0, " DCSB0[%d]=0x%08x reg: F2x%x\n", cs, *base0, reg0); if (pvt->fam == 0xf) continue; if (!amd64_read_dct_pci_cfg(pvt, 1, reg0, base1)) edac_dbg(0, " DCSB1[%d]=0x%08x reg: F2x%x\n", cs, *base1, (pvt->fam == 0x10) ? reg1 : reg0); } for_each_chip_select_mask(cs, 0, pvt) { int reg0 = DCSM0 + (cs * 4); int reg1 = DCSM1 + (cs * 4); u32 *mask0 = &pvt->csels[0].csmasks[cs]; u32 *mask1 = &pvt->csels[1].csmasks[cs]; if (!amd64_read_dct_pci_cfg(pvt, 0, reg0, mask0)) edac_dbg(0, " DCSM0[%d]=0x%08x reg: F2x%x\n", cs, *mask0, reg0); if (pvt->fam == 0xf) continue; if (!amd64_read_dct_pci_cfg(pvt, 1, reg0, mask1)) edac_dbg(0, " DCSM1[%d]=0x%08x reg: F2x%x\n", cs, *mask1, (pvt->fam == 0x10) ? reg1 : reg0); } } static void umc_determine_memory_type(struct amd64_pvt *pvt) { struct amd64_umc *umc; u32 i; for_each_umc(i) { umc = &pvt->umc[i]; if (!(umc->sdp_ctrl & UMC_SDP_INIT)) { umc->dram_type = MEM_EMPTY; continue; } /* * Check if the system supports the "DDR Type" field in UMC Config * and has DDR5 DIMMs in use. */ if (pvt->flags.zn_regs_v2 && ((umc->umc_cfg & GENMASK(2, 0)) == 0x1)) { if (umc->dimm_cfg & BIT(5)) umc->dram_type = MEM_LRDDR5; else if (umc->dimm_cfg & BIT(4)) umc->dram_type = MEM_RDDR5; else umc->dram_type = MEM_DDR5; } else { if (umc->dimm_cfg & BIT(5)) umc->dram_type = MEM_LRDDR4; else if (umc->dimm_cfg & BIT(4)) umc->dram_type = MEM_RDDR4; else umc->dram_type = MEM_DDR4; } edac_dbg(1, " UMC%d DIMM type: %s\n", i, edac_mem_types[umc->dram_type]); } } static void dct_determine_memory_type(struct amd64_pvt *pvt) { u32 dram_ctrl, dcsm; switch (pvt->fam) { case 0xf: if (pvt->ext_model >= K8_REV_F) goto ddr3; pvt->dram_type = (pvt->dclr0 & BIT(18)) ? MEM_DDR : MEM_RDDR; return; case 0x10: if (pvt->dchr0 & DDR3_MODE) goto ddr3; pvt->dram_type = (pvt->dclr0 & BIT(16)) ? MEM_DDR2 : MEM_RDDR2; return; case 0x15: if (pvt->model < 0x60) goto ddr3; /* * Model 0x60h needs special handling: * * We use a Chip Select value of '0' to obtain dcsm. * Theoretically, it is possible to populate LRDIMMs of different * 'Rank' value on a DCT. But this is not the common case. So, * it's reasonable to assume all DIMMs are going to be of same * 'type' until proven otherwise. */ amd64_read_dct_pci_cfg(pvt, 0, DRAM_CONTROL, &dram_ctrl); dcsm = pvt->csels[0].csmasks[0]; if (((dram_ctrl >> 8) & 0x7) == 0x2) pvt->dram_type = MEM_DDR4; else if (pvt->dclr0 & BIT(16)) pvt->dram_type = MEM_DDR3; else if (dcsm & 0x3) pvt->dram_type = MEM_LRDDR3; else pvt->dram_type = MEM_RDDR3; return; case 0x16: goto ddr3; default: WARN(1, KERN_ERR "%s: Family??? 0x%x\n", __func__, pvt->fam); pvt->dram_type = MEM_EMPTY; } edac_dbg(1, " DIMM type: %s\n", edac_mem_types[pvt->dram_type]); return; ddr3: pvt->dram_type = (pvt->dclr0 & BIT(16)) ? MEM_DDR3 : MEM_RDDR3; } /* On F10h and later ErrAddr is MC4_ADDR[47:1] */ static u64 get_error_address(struct amd64_pvt *pvt, struct mce *m) { u16 mce_nid = topology_die_id(m->extcpu); struct mem_ctl_info *mci; u8 start_bit = 1; u8 end_bit = 47; u64 addr; mci = edac_mc_find(mce_nid); if (!mci) return 0; pvt = mci->pvt_info; if (pvt->fam == 0xf) { start_bit = 3; end_bit = 39; } addr = m->addr & GENMASK_ULL(end_bit, start_bit); /* * Erratum 637 workaround */ if (pvt->fam == 0x15) { u64 cc6_base, tmp_addr; u32 tmp; u8 intlv_en; if ((addr & GENMASK_ULL(47, 24)) >> 24 != 0x00fdf7) return addr; amd64_read_pci_cfg(pvt->F1, DRAM_LOCAL_NODE_LIM, &tmp); intlv_en = tmp >> 21 & 0x7; /* add [47:27] + 3 trailing bits */ cc6_base = (tmp & GENMASK_ULL(20, 0)) << 3; /* reverse and add DramIntlvEn */ cc6_base |= intlv_en ^ 0x7; /* pin at [47:24] */ cc6_base <<= 24; if (!intlv_en) 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_ULL(23, 12)) << __fls(intlv_en + 1); /* OR DramIntlvSel into bits [14:12] */ tmp_addr |= (tmp & GENMASK_ULL(23, 21)) >> 9; /* add remaining [11:0] bits from original MC4_ADDR */ tmp_addr |= addr & GENMASK_ULL(11, 0); return cc6_base | tmp_addr; } return addr; } static struct pci_dev *pci_get_related_function(unsigned int vendor, unsigned int device, struct pci_dev *related) { struct pci_dev *dev = NULL; while ((dev = pci_get_device(vendor, device, dev))) { if (pci_domain_nr(dev->bus) == pci_domain_nr(related->bus) && (dev->bus->number == related->bus->number) && (PCI_SLOT(dev->devfn) == PCI_SLOT(related->devfn))) break; } return dev; } static void read_dram_base_limit_regs(struct amd64_pvt *pvt, unsigned range) { struct amd_northbridge *nb; struct pci_dev *f1 = NULL; unsigned int pci_func; int off = range << 3; u32 llim; amd64_read_pci_cfg(pvt->F1, DRAM_BASE_LO + off, &pvt->ranges[range].base.lo); amd64_read_pci_cfg(pvt->F1, DRAM_LIMIT_LO + off, &pvt->ranges[range].lim.lo); if (pvt->fam == 0xf) return; if (!dram_rw(pvt, range)) return; amd64_read_pci_cfg(pvt->F1, DRAM_BASE_HI + off, &pvt->ranges[range].base.hi); amd64_read_pci_cfg(pvt->F1, DRAM_LIMIT_HI + off, &pvt->ranges[range].lim.hi); /* F15h: factor in CC6 save area by reading dst node's limit reg */ if (pvt->fam != 0x15) return; nb = node_to_amd_nb(dram_dst_node(pvt, range)); if (WARN_ON(!nb)) return; if (pvt->model == 0x60) pci_func = PCI_DEVICE_ID_AMD_15H_M60H_NB_F1; else if (pvt->model == 0x30) pci_func = PCI_DEVICE_ID_AMD_15H_M30H_NB_F1; else pci_func = PCI_DEVICE_ID_AMD_15H_NB_F1; f1 = pci_get_related_function(nb->misc->vendor, pci_func, nb->misc); if (WARN_ON(!f1)) return; amd64_read_pci_cfg(f1, DRAM_LOCAL_NODE_LIM, &llim); 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_ULL(7, 0); /* [47:40] */ pvt->ranges[range].lim.hi |= llim >> 13; pci_dev_put(f1); } static void k8_map_sysaddr_to_csrow(struct mem_ctl_info *mci, u64 sys_addr, struct err_info *err) { struct amd64_pvt *pvt = mci->pvt_info; error_address_to_page_and_offset(sys_addr, err); /* * Find out which node the error address belongs to. This may be * different from the node that detected the error. */ err->src_mci = find_mc_by_sys_addr(mci, sys_addr); if (!err->src_mci) { amd64_mc_err(mci, "failed to map error addr 0x%lx to a node\n", (unsigned long)sys_addr); err->err_code = ERR_NODE; return; } /* Now map the sys_addr to a CSROW */ err->csrow = sys_addr_to_csrow(err->src_mci, sys_addr); if (err->csrow < 0) { err->err_code = ERR_CSROW; return; } /* CHIPKILL enabled */ if (pvt->nbcfg & NBCFG_CHIPKILL) { err->channel = get_channel_from_ecc_syndrome(mci, err->syndrome); if (err->channel < 0) { /* * Syndrome didn't map, so we don't know which of the * 2 DIMMs is in error. So we need to ID 'both' of them * as suspect. */ amd64_mc_warn(err->src_mci, "unknown syndrome 0x%04x - " "possible error reporting race\n", err->syndrome); err->err_code = ERR_CHANNEL; return; } } else { /* * non-chipkill ecc mode * * The k8 documentation is unclear about how to determine the * channel number when using non-chipkill memory. This method * was obtained from email communication with someone at AMD. * (Wish the email was placed in this comment - norsk) */ err->channel = ((sys_addr & BIT(3)) != 0); } } static int ddr2_cs_size(unsigned i, bool dct_width) { unsigned shift = 0; if (i <= 2) shift = i; else if (!(i & 0x1)) shift = i >> 1; else shift = (i + 1) >> 1; return 128 << (shift + !!dct_width); } static int k8_dbam_to_chip_select(struct amd64_pvt *pvt, u8 dct, unsigned cs_mode, int cs_mask_nr) { u32 dclr = dct ? pvt->dclr1 : pvt->dclr0; if (pvt->ext_model >= K8_REV_F) { WARN_ON(cs_mode > 11); return ddr2_cs_size(cs_mode, dclr & WIDTH_128); } else if (pvt->ext_model >= K8_REV_D) { unsigned diff; WARN_ON(cs_mode > 10); /* * the below calculation, besides trying to win an obfuscated C * contest, maps cs_mode values to DIMM chip select sizes. The * mappings are: * * cs_mode CS size (mb) * ======= ============ * 0 32 * 1 64 * 2 128 * 3 128 * 4 256 * 5 512 * 6 256 * 7 512 * 8 1024 * 9 1024 * 10 2048 * * Basically, it calculates a value with which to shift the * smallest CS size of 32MB. * * ddr[23]_cs_size have a similar purpose. */ diff = cs_mode/3 + (unsigned)(cs_mode > 5); return 32 << (cs_mode - diff); } else { WARN_ON(cs_mode > 6); return 32 << cs_mode; } } static int ddr3_cs_size(unsigned i, bool dct_width) { unsigned shift = 0; int cs_size = 0; if (i == 0 || i == 3 || i == 4) cs_size = -1; else if (i <= 2) shift = i; else if (i == 12) shift = 7; else if (!(i & 0x1)) shift = i >> 1; else shift = (i + 1) >> 1; if (cs_size != -1) cs_size = (128 * (1 << !!dct_width)) << shift; return cs_size; } static int ddr3_lrdimm_cs_size(unsigned i, unsigned rank_multiply) { unsigned shift = 0; int cs_size = 0; if (i < 4 || i == 6) cs_size = -1; else if (i == 12) shift = 7; else if (!(i & 0x1)) shift = i >> 1; else shift = (i + 1) >> 1; if (cs_size != -1) cs_size = rank_multiply * (128 << shift); return cs_size; } static int ddr4_cs_size(unsigned i) { int cs_size = 0; if (i == 0) cs_size = -1; else if (i == 1) cs_size = 1024; else /* Min cs_size = 1G */ cs_size = 1024 * (1 << (i >> 1)); return cs_size; } static int f10_dbam_to_chip_select(struct amd64_pvt *pvt, u8 dct, unsigned cs_mode, int cs_mask_nr) { u32 dclr = dct ? pvt->dclr1 : pvt->dclr0; WARN_ON(cs_mode > 11); if (pvt->dchr0 & DDR3_MODE || pvt->dchr1 & DDR3_MODE) return ddr3_cs_size(cs_mode, dclr & WIDTH_128); else return ddr2_cs_size(cs_mode, dclr & WIDTH_128); } /* * F15h supports only 64bit DCT interfaces */ static int f15_dbam_to_chip_select(struct amd64_pvt *pvt, u8 dct, unsigned cs_mode, int cs_mask_nr) { WARN_ON(cs_mode > 12); return ddr3_cs_size(cs_mode, false); } /* F15h M60h supports DDR4 mapping as well.. */ static int f15_m60h_dbam_to_chip_select(struct amd64_pvt *pvt, u8 dct, unsigned cs_mode, int cs_mask_nr) { int cs_size; u32 dcsm = pvt->csels[dct].csmasks[cs_mask_nr]; WARN_ON(cs_mode > 12); if (pvt->dram_type == MEM_DDR4) { if (cs_mode > 9) return -1; cs_size = ddr4_cs_size(cs_mode); } else if (pvt->dram_type == MEM_LRDDR3) { unsigned rank_multiply = dcsm & 0xf; if (rank_multiply == 3) rank_multiply = 4; cs_size = ddr3_lrdimm_cs_size(cs_mode, rank_multiply); } else { /* Minimum cs size is 512mb for F15hM60h*/ if (cs_mode == 0x1) return -1; cs_size = ddr3_cs_size(cs_mode, false); } return cs_size; } /* * F16h and F15h model 30h have only limited cs_modes. */ static int f16_dbam_to_chip_select(struct amd64_pvt *pvt, u8 dct, unsigned cs_mode, int cs_mask_nr) { WARN_ON(cs_mode > 12); if (cs_mode == 6 || cs_mode == 8 || cs_mode == 9 || cs_mode == 12) return -1; else return ddr3_cs_size(cs_mode, false); } static void read_dram_ctl_register(struct amd64_pvt *pvt) { if (pvt->fam == 0xf) return; if (!amd64_read_pci_cfg(pvt->F2, DCT_SEL_LO, &pvt->dct_sel_lo)) { edac_dbg(0, "F2x110 (DCTSelLow): 0x%08x, High range addrs at: 0x%x\n", pvt->dct_sel_lo, dct_sel_baseaddr(pvt)); edac_dbg(0, " DCTs operate in %s mode\n", (dct_ganging_enabled(pvt) ? "ganged" : "unganged")); if (!dct_ganging_enabled(pvt)) edac_dbg(0, " Address range split per DCT: %s\n", (dct_high_range_enabled(pvt) ? "yes" : "no")); edac_dbg(0, " data interleave for ECC: %s, DRAM cleared since last warm reset: %s\n", (dct_data_intlv_enabled(pvt) ? "enabled" : "disabled"), (dct_memory_cleared(pvt) ? "yes" : "no")); edac_dbg(0, " channel interleave: %s, " "interleave bits selector: 0x%x\n", (dct_interleave_enabled(pvt) ? "enabled" : "disabled"), dct_sel_interleave_addr(pvt)); } amd64_read_pci_cfg(pvt->F2, DCT_SEL_HI, &pvt->dct_sel_hi); } /* * Determine channel (DCT) based on the interleaving mode (see F15h M30h BKDG, * 2.10.12 Memory Interleaving Modes). */ static u8 f15_m30h_determine_channel(struct amd64_pvt *pvt, u64 sys_addr, u8 intlv_en, int num_dcts_intlv, u32 dct_sel) { u8 channel = 0; u8 select; if (!(intlv_en)) return (u8)(dct_sel); if (num_dcts_intlv == 2) { select = (sys_addr >> 8) & 0x3; channel = select ? 0x3 : 0; } 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; } /* * Determine channel (DCT) based on the interleaving mode: F10h BKDG, 2.8.9 Memory * Interleaving Modes. */ static u8 f1x_determine_channel(struct amd64_pvt *pvt, u64 sys_addr, bool hi_range_sel, u8 intlv_en) { u8 dct_sel_high = (pvt->dct_sel_lo >> 1) & 1; if (dct_ganging_enabled(pvt)) return 0; if (hi_range_sel) return dct_sel_high; /* * see F2x110[DctSelIntLvAddr] - channel interleave mode */ if (dct_interleave_enabled(pvt)) { u8 intlv_addr = dct_sel_interleave_addr(pvt); /* return DCT select function: 0=DCT0, 1=DCT1 */ if (!intlv_addr) return sys_addr >> 6 & 1; if (intlv_addr & 0x2) { u8 shift = intlv_addr & 0x1 ? 9 : 6; u32 temp = hweight_long((u32) ((sys_addr >> 16) & 0x1F)) & 1; return ((sys_addr >> shift) & 1) ^ temp; } if (intlv_addr & 0x4) { u8 shift = intlv_addr & 0x1 ? 9 : 8; return (sys_addr >> shift) & 1; } return (sys_addr >> (12 + hweight8(intlv_en))) & 1; } if (dct_high_range_enabled(pvt)) return ~dct_sel_high & 1; return 0; } /* Convert the sys_addr to the normalized DCT address */ static u64 f1x_get_norm_dct_addr(struct amd64_pvt *pvt, u8 range, u64 sys_addr, bool hi_rng, u32 dct_sel_base_addr) { u64 chan_off; u64 dram_base = get_dram_base(pvt, range); u64 hole_off = f10_dhar_offset(pvt); u64 dct_sel_base_off = (u64)(pvt->dct_sel_hi & 0xFFFFFC00) << 16; if (hi_rng) { /* * if * base address of high range is below 4Gb * (bits [47:27] at [31:11]) * DRAM address space on this DCT is hoisted above 4Gb && * sys_addr > 4Gb * * remove hole offset from sys_addr * else * remove high range offset from sys_addr */ if ((!(dct_sel_base_addr >> 16) || dct_sel_base_addr < dhar_base(pvt)) && dhar_valid(pvt) && (sys_addr >= BIT_64(32))) chan_off = hole_off; else chan_off = dct_sel_base_off; } else { /* * if * we have a valid hole && * sys_addr > 4Gb * * remove hole * else * remove dram base to normalize to DCT address */ if (dhar_valid(pvt) && (sys_addr >= BIT_64(32))) chan_off = hole_off; else chan_off = dram_base; } return (sys_addr & GENMASK_ULL(47,6)) - (chan_off & GENMASK_ULL(47,23)); } /* * checks if the csrow passed in is marked as SPARED, if so returns the new * spare row */ static int f10_process_possible_spare(struct amd64_pvt *pvt, u8 dct, int csrow) { int tmp_cs; if (online_spare_swap_done(pvt, dct) && csrow == online_spare_bad_dramcs(pvt, dct)) { for_each_chip_select(tmp_cs, dct, pvt) { if (chip_select_base(tmp_cs, dct, pvt) & 0x2) { csrow = tmp_cs; break; } } } return csrow; } /* * Iterate over the DRAM DCT "base" and "mask" registers looking for a * SystemAddr match on the specified 'ChannelSelect' and 'NodeID' * * Return: * -EINVAL: NOT FOUND * 0..csrow = Chip-Select Row */ static int f1x_lookup_addr_in_dct(u64 in_addr, u8 nid, u8 dct) { struct mem_ctl_info *mci; struct amd64_pvt *pvt; u64 cs_base, cs_mask; int cs_found = -EINVAL; int csrow; mci = edac_mc_find(nid); if (!mci) return cs_found; pvt = mci->pvt_info; edac_dbg(1, "input addr: 0x%llx, DCT: %d\n", in_addr, dct); for_each_chip_select(csrow, dct, pvt) { if (!csrow_enabled(csrow, dct, pvt)) continue; get_cs_base_and_mask(pvt, csrow, dct, &cs_base, &cs_mask); edac_dbg(1, " CSROW=%d CSBase=0x%llx CSMask=0x%llx\n", csrow, cs_base, cs_mask); cs_mask = ~cs_mask; edac_dbg(1, " (InputAddr & ~CSMask)=0x%llx (CSBase & ~CSMask)=0x%llx\n", (in_addr & cs_mask), (cs_base & cs_mask)); if ((in_addr & cs_mask) == (cs_base & cs_mask)) { if (pvt->fam == 0x15 && pvt->model >= 0x30) { cs_found = csrow; break; } cs_found = f10_process_possible_spare(pvt, dct, csrow); edac_dbg(1, " MATCH csrow=%d\n", cs_found); break; } } return cs_found; } /* * See F2x10C. Non-interleaved graphics framebuffer memory under the 16G is * swapped with a region located at the bottom of memory so that the GPU can use * the interleaved region and thus two channels. */ static u64 f1x_swap_interleaved_region(struct amd64_pvt *pvt, u64 sys_addr) { u32 swap_reg, swap_base, swap_limit, rgn_size, tmp_addr; if (pvt->fam == 0x10) { /* only revC3 and revE have that feature */ if (pvt->model < 4 || (pvt->model < 0xa && pvt->stepping < 3)) return sys_addr; } amd64_read_pci_cfg(pvt->F2, SWAP_INTLV_REG, &swap_reg); if (!(swap_reg & 0x1)) return sys_addr; swap_base = (swap_reg >> 3) & 0x7f; swap_limit = (swap_reg >> 11) & 0x7f; rgn_size = (swap_reg >> 20) & 0x7f; tmp_addr = sys_addr >> 27; if (!(sys_addr >> 34) && (((tmp_addr >= swap_base) && (tmp_addr <= swap_limit)) || (tmp_addr < rgn_size))) return sys_addr ^ (u64)swap_base << 27; return sys_addr; } /* For a given @dram_range, check if @sys_addr falls within it. */ static int f1x_match_to_this_node(struct amd64_pvt *pvt, unsigned range, u64 sys_addr, int *chan_sel) { int cs_found = -EINVAL; u64 chan_addr; u32 dct_sel_base; u8 channel; bool high_range = false; u8 node_id = dram_dst_node(pvt, range); u8 intlv_en = dram_intlv_en(pvt, range); u32 intlv_sel = dram_intlv_sel(pvt, range); edac_dbg(1, "(range %d) SystemAddr= 0x%llx Limit=0x%llx\n", range, sys_addr, get_dram_limit(pvt, range)); if (dhar_valid(pvt) && dhar_base(pvt) <= sys_addr && sys_addr < BIT_64(32)) { amd64_warn("Huh? Address is in the MMIO hole: 0x%016llx\n", sys_addr); return -EINVAL; } if (intlv_en && (intlv_sel != ((sys_addr >> 12) & intlv_en))) return -EINVAL; sys_addr = f1x_swap_interleaved_region(pvt, sys_addr); dct_sel_base = dct_sel_baseaddr(pvt); /* * check whether addresses >= DctSelBaseAddr[47:27] are to be used to * select between DCT0 and DCT1. */ if (dct_high_range_enabled(pvt) && !dct_ganging_enabled(pvt) && ((sys_addr >> 27) >= (dct_sel_base >> 11))) high_range = true; channel = f1x_determine_channel(pvt, sys_addr, high_range, intlv_en); chan_addr = f1x_get_norm_dct_addr(pvt, range, sys_addr, high_range, dct_sel_base); /* Remove node interleaving, see F1x120 */ if (intlv_en) chan_addr = ((chan_addr >> (12 + hweight8(intlv_en))) << 12) | (chan_addr & 0xfff); /* remove channel interleave */ if (dct_interleave_enabled(pvt) && !dct_high_range_enabled(pvt) && !dct_ganging_enabled(pvt)) { if (dct_sel_interleave_addr(pvt) != 1) { if (dct_sel_interleave_addr(pvt) == 0x3) /* hash 9 */ chan_addr = ((chan_addr >> 10) << 9) | (chan_addr & 0x1ff); else /* A[6] or hash 6 */ chan_addr = ((chan_addr >> 7) << 6) | (chan_addr & 0x3f); } else /* A[12] */ chan_addr = ((chan_addr >> 13) << 12) | (chan_addr & 0xfff); } edac_dbg(1, " Normalized DCT addr: 0x%llx\n", chan_addr); cs_found = f1x_lookup_addr_in_dct(chan_addr, node_id, channel); if (cs_found >= 0) *chan_sel = channel; return cs_found; } static int f15_m30h_match_to_this_node(struct amd64_pvt *pvt, unsigned range, u64 sys_addr, int *chan_sel) { int cs_found = -EINVAL; int num_dcts_intlv = 0; u64 chan_addr, chan_offset; u64 dct_base, dct_limit; u32 dct_cont_base_reg, dct_cont_limit_reg, tmp; u8 channel, alias_channel, leg_mmio_hole, dct_sel, dct_offset_en; u64 dhar_offset = f10_dhar_offset(pvt); u8 intlv_addr = dct_sel_interleave_addr(pvt); u8 node_id = dram_dst_node(pvt, range); u8 intlv_en = dram_intlv_en(pvt, range); amd64_read_pci_cfg(pvt->F1, DRAM_CONT_BASE, &dct_cont_base_reg); amd64_read_pci_cfg(pvt->F1, DRAM_CONT_LIMIT, &dct_cont_limit_reg); dct_offset_en = (u8) ((dct_cont_base_reg >> 3) & BIT(0)); dct_sel = (u8) ((dct_cont_base_reg >> 4) & 0x7); edac_dbg(1, "(range %d) SystemAddr= 0x%llx Limit=0x%llx\n", range, sys_addr, get_dram_limit(pvt, range)); if (!(get_dram_base(pvt, range) <= sys_addr) && !(get_dram_limit(pvt, range) >= sys_addr)) return -EINVAL; if (dhar_valid(pvt) && dhar_base(pvt) <= sys_addr && sys_addr < BIT_64(32)) { amd64_warn("Huh? Address is in the MMIO hole: 0x%016llx\n", sys_addr); return -EINVAL; } /* Verify sys_addr is within DCT Range. */ dct_base = (u64) dct_sel_baseaddr(pvt); dct_limit = (dct_cont_limit_reg >> 11) & 0x1FFF; if (!(dct_cont_base_reg & BIT(0)) && !(dct_base <= (sys_addr >> 27) && dct_limit >= (sys_addr >> 27))) return -EINVAL; /* Verify number of dct's that participate in channel interleaving. */ num_dcts_intlv = (int) hweight8(intlv_en); if (!(num_dcts_intlv % 2 == 0) || (num_dcts_intlv > 4)) return -EINVAL; if (pvt->model >= 0x60) channel = f1x_determine_channel(pvt, sys_addr, false, intlv_en); else channel = f15_m30h_determine_channel(pvt, sys_addr, intlv_en, num_dcts_intlv, dct_sel); /* Verify we stay within the MAX number of channels allowed */ if (channel > 3) return -EINVAL; leg_mmio_hole = (u8) (dct_cont_base_reg >> 1 & BIT(0)); /* Get normalized DCT addr */ if (leg_mmio_hole && (sys_addr >= BIT_64(32))) chan_offset = dhar_offset; else chan_offset = dct_base << 27; chan_addr = sys_addr - chan_offset; /* remove channel interleave */ if (num_dcts_intlv == 2) { if (intlv_addr == 0x4) chan_addr = ((chan_addr >> 9) << 8) | (chan_addr & 0xff); else if (intlv_addr == 0x5) chan_addr = ((chan_addr >> 10) << 9) | (chan_addr & 0x1ff); else return -EINVAL; } else if (num_dcts_intlv == 4) { if (intlv_addr == 0x4) chan_addr = ((chan_addr >> 10) << 8) | (chan_addr & 0xff); else if (intlv_addr == 0x5) chan_addr = ((chan_addr >> 11) << 9) | (chan_addr & 0x1ff); else return -EINVAL; } if (dct_offset_en) { amd64_read_pci_cfg(pvt->F1, DRAM_CONT_HIGH_OFF + (int) channel * 4, &tmp); chan_addr += (u64) ((tmp >> 11) & 0xfff) << 27; } f15h_select_dct(pvt, channel); edac_dbg(1, " Normalized DCT addr: 0x%llx\n", chan_addr); /* * Find Chip select: * if channel = 3, then alias it to 1. This is because, in F15 M30h, * there is support for 4 DCT's, but only 2 are currently functional. * They are DCT0 and DCT3. But we have read all registers of DCT3 into * pvt->csels[1]. So we need to use '1' here to get correct info. * Refer F15 M30h BKDG Section 2.10 and 2.10.3 for clarifications. */ alias_channel = (channel == 3) ? 1 : channel; cs_found = f1x_lookup_addr_in_dct(chan_addr, node_id, alias_channel); if (cs_found >= 0) *chan_sel = alias_channel; return cs_found; } static int f1x_translate_sysaddr_to_cs(struct amd64_pvt *pvt, u64 sys_addr, int *chan_sel) { int cs_found = -EINVAL; unsigned range; for (range = 0; range < DRAM_RANGES; range++) { if (!dram_rw(pvt, range)) continue; if (pvt->fam == 0x15 && pvt->model >= 0x30) cs_found = f15_m30h_match_to_this_node(pvt, range, sys_addr, chan_sel); else if ((get_dram_base(pvt, range) <= sys_addr) && (get_dram_limit(pvt, range) >= sys_addr)) { cs_found = f1x_match_to_this_node(pvt, range, sys_addr, chan_sel); if (cs_found >= 0) break; } } return cs_found; } /* * For reference see "2.8.5 Routing DRAM Requests" in F10 BKDG. This code maps * a @sys_addr to NodeID, DCT (channel) and chip select (CSROW). * * The @sys_addr is usually an error address received from the hardware * (MCX_ADDR). */ static void f1x_map_sysaddr_to_csrow(struct mem_ctl_info *mci, u64 sys_addr, struct err_info *err) { struct amd64_pvt *pvt = mci->pvt_info; error_address_to_page_and_offset(sys_addr, err); err->csrow = f1x_translate_sysaddr_to_cs(pvt, sys_addr, &err->channel); if (err->csrow < 0) { err->err_code = ERR_CSROW; return; } /* * We need the syndromes for channel detection only when we're * ganged. Otherwise @chan should already contain the channel at * this point. */ if (dct_ganging_enabled(pvt)) err->channel = get_channel_from_ecc_syndrome(mci, err->syndrome); } /* * These are tables of eigenvectors (one per line) which can be used for the * construction of the syndrome tables. The modified syndrome search algorithm * uses those to find the symbol in error and thus the DIMM. * * Algorithm courtesy of Ross LaFetra from AMD. */ static const u16 x4_vectors[] = { 0x2f57, 0x1afe, 0x66cc, 0xdd88, 0x11eb, 0x3396, 0x7f4c, 0xeac8, 0x0001, 0x0002, 0x0004, 0x0008, 0x1013, 0x3032, 0x4044, 0x8088, 0x106b, 0x30d6, 0x70fc, 0xe0a8, 0x4857, 0xc4fe, 0x13cc, 0x3288, 0x1ac5, 0x2f4a, 0x5394, 0xa1e8, 0x1f39, 0x251e, 0xbd6c, 0x6bd8, 0x15c1, 0x2a42, 0x89ac, 0x4758, 0x2b03, 0x1602, 0x4f0c, 0xca08, 0x1f07, 0x3a0e, 0x6b04, 0xbd08, 0x8ba7, 0x465e, 0x244c, 0x1cc8, 0x2b87, 0x164e, 0x642c, 0xdc18, 0x40b9, 0x80de, 0x1094, 0x20e8, 0x27db, 0x1eb6, 0x9dac, 0x7b58, 0x11c1, 0x2242, 0x84ac, 0x4c58, 0x1be5, 0x2d7a, 0x5e34, 0xa718, 0x4b39, 0x8d1e, 0x14b4, 0x28d8, 0x4c97, 0xc87e, 0x11fc, 0x33a8, 0x8e97, 0x497e, 0x2ffc, 0x1aa8, 0x16b3, 0x3d62, 0x4f34, 0x8518, 0x1e2f, 0x391a, 0x5cac, 0xf858, 0x1d9f, 0x3b7a, 0x572c, 0xfe18, 0x15f5, 0x2a5a, 0x5264, 0xa3b8, 0x1dbb, 0x3b66, 0x715c, 0xe3f8, 0x4397, 0xc27e, 0x17fc, 0x3ea8, 0x1617, 0x3d3e, 0x6464, 0xb8b8, 0x23ff, 0x12aa, 0xab6c, 0x56d8, 0x2dfb, 0x1ba6, 0x913c, 0x7328, 0x185d, 0x2ca6, 0x7914, 0x9e28, 0x171b, 0x3e36, 0x7d7c, 0xebe8, 0x4199, 0x82ee, 0x19f4, 0x2e58, 0x4807, 0xc40e, 0x130c, 0x3208, 0x1905, 0x2e0a, 0x5804, 0xac08, 0x213f, 0x132a, 0xadfc, 0x5ba8, 0x19a9, 0x2efe, 0xb5cc, 0x6f88, }; static const u16 x8_vectors[] = { 0x0145, 0x028a, 0x2374, 0x43c8, 0xa1f0, 0x0520, 0x0a40, 0x1480, 0x0211, 0x0422, 0x0844, 0x1088, 0x01b0, 0x44e0, 0x23c0, 0xed80, 0x1011, 0x0116, 0x022c, 0x0458, 0x08b0, 0x8c60, 0x2740, 0x4e80, 0x0411, 0x0822, 0x1044, 0x0158, 0x02b0, 0x2360, 0x46c0, 0xab80, 0x0811, 0x1022, 0x012c, 0x0258, 0x04b0, 0x4660, 0x8cc0, 0x2780, 0x2071, 0x40e2, 0xa0c4, 0x0108, 0x0210, 0x0420, 0x0840, 0x1080, 0x4071, 0x80e2, 0x0104, 0x0208, 0x0410, 0x0820, 0x1040, 0x2080, 0x8071, 0x0102, 0x0204, 0x0408, 0x0810, 0x1020, 0x2040, 0x4080, 0x019d, 0x03d6, 0x136c, 0x2198, 0x50b0, 0xb2e0, 0x0740, 0x0e80, 0x0189, 0x03ea, 0x072c, 0x0e58, 0x1cb0, 0x56e0, 0x37c0, 0xf580, 0x01fd, 0x0376, 0x06ec, 0x0bb8, 0x1110, 0x2220, 0x4440, 0x8880, 0x0163, 0x02c6, 0x1104, 0x0758, 0x0eb0, 0x2be0, 0x6140, 0xc280, 0x02fd, 0x01c6, 0x0b5c, 0x1108, 0x07b0, 0x25a0, 0x8840, 0x6180, 0x0801, 0x012e, 0x025c, 0x04b8, 0x1370, 0x26e0, 0x57c0, 0xb580, 0x0401, 0x0802, 0x015c, 0x02b8, 0x22b0, 0x13e0, 0x7140, 0xe280, 0x0201, 0x0402, 0x0804, 0x01b8, 0x11b0, 0x31a0, 0x8040, 0x7180, 0x0101, 0x0202, 0x0404, 0x0808, 0x1010, 0x2020, 0x4040, 0x8080, 0x0001, 0x0002, 0x0004, 0x0008, 0x0010, 0x0020, 0x0040, 0x0080, 0x0100, 0x0200, 0x0400, 0x0800, 0x1000, 0x2000, 0x4000, 0x8000, }; static int decode_syndrome(u16 syndrome, const u16 *vectors, unsigned num_vecs, unsigned v_dim) { unsigned int i, err_sym; for (err_sym = 0; err_sym < num_vecs / v_dim; err_sym++) { u16 s = syndrome; unsigned v_idx = err_sym * v_dim; unsigned v_end = (err_sym + 1) * v_dim; /* walk over all 16 bits of the syndrome */ for (i = 1; i < (1U << 16); i <<= 1) { /* if bit is set in that eigenvector... */ if (v_idx < v_end && vectors[v_idx] & i) { u16 ev_comp = vectors[v_idx++]; /* ... and bit set in the modified syndrome, */ if (s & i) { /* remove it. */ s ^= ev_comp; if (!s) return err_sym; } } else if (s & i) /* can't get to zero, move to next symbol */ break; } } edac_dbg(0, "syndrome(%x) not found\n", syndrome); return -1; } static int map_err_sym_to_channel(int err_sym, int sym_size) { if (sym_size == 4) switch (err_sym) { case 0x20: case 0x21: return 0; case 0x22: case 0x23: return 1; default: return err_sym >> 4; } /* x8 symbols */ else switch (err_sym) { /* imaginary bits not in a DIMM */ case 0x10: WARN(1, KERN_ERR "Invalid error symbol: 0x%x\n", err_sym); return -1; case 0x11: return 0; case 0x12: return 1; default: return err_sym >> 3; } return -1; } static int get_channel_from_ecc_syndrome(struct mem_ctl_info *mci, u16 syndrome) { struct amd64_pvt *pvt = mci->pvt_info; int err_sym = -1; if (pvt->ecc_sym_sz == 8) err_sym = decode_syndrome(syndrome, x8_vectors, ARRAY_SIZE(x8_vectors), pvt->ecc_sym_sz); else if (pvt->ecc_sym_sz == 4) err_sym = decode_syndrome(syndrome, x4_vectors, ARRAY_SIZE(x4_vectors), pvt->ecc_sym_sz); else { amd64_warn("Illegal syndrome type: %u\n", pvt->ecc_sym_sz); return err_sym; } return map_err_sym_to_channel(err_sym, pvt->ecc_sym_sz); } static void __log_ecc_error(struct mem_ctl_info *mci, struct err_info *err, u8 ecc_type) { enum hw_event_mc_err_type err_type; const char *string; if (ecc_type == 2) err_type = HW_EVENT_ERR_CORRECTED; else if (ecc_type == 1) err_type = HW_EVENT_ERR_UNCORRECTED; else if (ecc_type == 3) err_type = HW_EVENT_ERR_DEFERRED; else { WARN(1, "Something is rotten in the state of Denmark.\n"); return; } switch (err->err_code) { case DECODE_OK: string = ""; break; case ERR_NODE: string = "Failed to map error addr to a node"; break; case ERR_CSROW: string = "Failed to map error addr to a csrow"; break; case ERR_CHANNEL: string = "Unknown syndrome - possible error reporting race"; break; case ERR_SYND: string = "MCA_SYND not valid - unknown syndrome and csrow"; break; case ERR_NORM_ADDR: string = "Cannot decode normalized address"; break; default: string = "WTF error"; break; } edac_mc_handle_error(err_type, mci, 1, err->page, err->offset, err->syndrome, err->csrow, err->channel, -1, string, ""); } static inline void decode_bus_error(int node_id, struct mce *m) { struct mem_ctl_info *mci; struct amd64_pvt *pvt; u8 ecc_type = (m->status >> 45) & 0x3; u8 xec = XEC(m->status, 0x1f); u16 ec = EC(m->status); u64 sys_addr; struct err_info err; mci = edac_mc_find(node_id); if (!mci) return; pvt = mci->pvt_info; /* Bail out early if this was an 'observed' error */ if (PP(ec) == NBSL_PP_OBS) return; /* Do only ECC errors */ if (xec && xec != F10_NBSL_EXT_ERR_ECC) return; memset(&err, 0, sizeof(err)); sys_addr = get_error_address(pvt, m); if (ecc_type == 2) err.syndrome = extract_syndrome(m->status); pvt->ops->map_sysaddr_to_csrow(mci, sys_addr, &err); __log_ecc_error(mci, &err, ecc_type); } /* * To find the UMC channel represented by this bank we need to match on its * instance_id. The instance_id of a bank is held in the lower 32 bits of its * IPID. * * Currently, we can derive the channel number by looking at the 6th nibble in * the instance_id. For example, instance_id=0xYXXXXX where Y is the channel * number. * * For DRAM ECC errors, the Chip Select number is given in bits [2:0] of * the MCA_SYND[ErrorInformation] field. */ static void umc_get_err_info(struct mce *m, struct err_info *err) { err->channel = (m->ipid & GENMASK(31, 0)) >> 20; err->csrow = m->synd & 0x7; } static void decode_umc_error(int node_id, struct mce *m) { u8 ecc_type = (m->status >> 45) & 0x3; struct mem_ctl_info *mci; struct amd64_pvt *pvt; struct err_info err; u64 sys_addr; node_id = fixup_node_id(node_id, m); mci = edac_mc_find(node_id); if (!mci) return; pvt = mci->pvt_info; memset(&err, 0, sizeof(err)); if (m->status & MCI_STATUS_DEFERRED) ecc_type = 3; if (!(m->status & MCI_STATUS_SYNDV)) { err.err_code = ERR_SYND; goto log_error; } if (ecc_type == 2) { u8 length = (m->synd >> 18) & 0x3f; if (length) err.syndrome = (m->synd >> 32) & GENMASK(length - 1, 0); else err.err_code = ERR_CHANNEL; } pvt->ops->get_err_info(m, &err); if (umc_normaddr_to_sysaddr(m->addr, pvt->mc_node_id, err.channel, &sys_addr)) { err.err_code = ERR_NORM_ADDR; goto log_error; } error_address_to_page_and_offset(sys_addr, &err); log_error: __log_ecc_error(mci, &err, ecc_type); } /* * Use pvt->F3 which contains the F3 CPU PCI device to get the related * F1 (AddrMap) and F2 (Dct) devices. Return negative value on error. */ static int reserve_mc_sibling_devs(struct amd64_pvt *pvt, u16 pci_id1, u16 pci_id2) { /* Reserve the ADDRESS MAP Device */ pvt->F1 = pci_get_related_function(pvt->F3->vendor, pci_id1, pvt->F3); if (!pvt->F1) { edac_dbg(1, "F1 not found: device 0x%x\n", pci_id1); return -ENODEV; } /* Reserve the DCT Device */ pvt->F2 = pci_get_related_function(pvt->F3->vendor, pci_id2, pvt->F3); if (!pvt->F2) { pci_dev_put(pvt->F1); pvt->F1 = NULL; edac_dbg(1, "F2 not found: device 0x%x\n", pci_id2); return -ENODEV; } if (!pci_ctl_dev) pci_ctl_dev = &pvt->F2->dev; edac_dbg(1, "F1: %s\n", pci_name(pvt->F1)); edac_dbg(1, "F2: %s\n", pci_name(pvt->F2)); edac_dbg(1, "F3: %s\n", pci_name(pvt->F3)); return 0; } static void determine_ecc_sym_sz(struct amd64_pvt *pvt) { pvt->ecc_sym_sz = 4; if (pvt->fam >= 0x10) { u32 tmp; amd64_read_pci_cfg(pvt->F3, EXT_NB_MCA_CFG, &tmp); /* F16h has only DCT0, so no need to read dbam1. */ if (pvt->fam != 0x16) amd64_read_dct_pci_cfg(pvt, 1, DBAM0, &pvt->dbam1); /* F10h, revD and later can do x8 ECC too. */ if ((pvt->fam > 0x10 || pvt->model > 7) && tmp & BIT(25)) pvt->ecc_sym_sz = 8; } } /* * Retrieve the hardware registers of the memory controller. */ static void umc_read_mc_regs(struct amd64_pvt *pvt) { u8 nid = pvt->mc_node_id; struct amd64_umc *umc; u32 i, umc_base; /* Read registers from each UMC */ for_each_umc(i) { umc_base = get_umc_base(i); umc = &pvt->umc[i]; amd_smn_read(nid, umc_base + get_umc_reg(pvt, UMCCH_DIMM_CFG), &umc->dimm_cfg); amd_smn_read(nid, umc_base + UMCCH_UMC_CFG, &umc->umc_cfg); amd_smn_read(nid, umc_base + UMCCH_SDP_CTRL, &umc->sdp_ctrl); amd_smn_read(nid, umc_base + UMCCH_ECC_CTRL, &umc->ecc_ctrl); amd_smn_read(nid, umc_base + UMCCH_UMC_CAP_HI, &umc->umc_cap_hi); } } /* * Retrieve the hardware registers of the memory controller (this includes the * 'Address Map' and 'Misc' device regs) */ static void dct_read_mc_regs(struct amd64_pvt *pvt) { unsigned int range; u64 msr_val; /* * Retrieve TOP_MEM and TOP_MEM2; no masking off of reserved bits since * those are Read-As-Zero. */ rdmsrl(MSR_K8_TOP_MEM1, pvt->top_mem); edac_dbg(0, " TOP_MEM: 0x%016llx\n", pvt->top_mem); /* Check first whether TOP_MEM2 is enabled: */ rdmsrl(MSR_AMD64_SYSCFG, msr_val); if (msr_val & BIT(21)) { rdmsrl(MSR_K8_TOP_MEM2, pvt->top_mem2); edac_dbg(0, " TOP_MEM2: 0x%016llx\n", pvt->top_mem2); } else { edac_dbg(0, " TOP_MEM2 disabled\n"); } amd64_read_pci_cfg(pvt->F3, NBCAP, &pvt->nbcap); read_dram_ctl_register(pvt); for (range = 0; range < DRAM_RANGES; range++) { u8 rw; /* read settings for this DRAM range */ read_dram_base_limit_regs(pvt, range); rw = dram_rw(pvt, range); if (!rw) continue; edac_dbg(1, " DRAM range[%d], base: 0x%016llx; limit: 0x%016llx\n", range, get_dram_base(pvt, range), get_dram_limit(pvt, range)); edac_dbg(1, " IntlvEn=%s; Range access: %s%s IntlvSel=%d DstNode=%d\n", dram_intlv_en(pvt, range) ? "Enabled" : "Disabled", (rw & 0x1) ? "R" : "-", (rw & 0x2) ? "W" : "-", dram_intlv_sel(pvt, range), dram_dst_node(pvt, range)); } amd64_read_pci_cfg(pvt->F1, DHAR, &pvt->dhar); amd64_read_dct_pci_cfg(pvt, 0, DBAM0, &pvt->dbam0); amd64_read_pci_cfg(pvt->F3, F10_ONLINE_SPARE, &pvt->online_spare); amd64_read_dct_pci_cfg(pvt, 0, DCLR0, &pvt->dclr0); amd64_read_dct_pci_cfg(pvt, 0, DCHR0, &pvt->dchr0); if (!dct_ganging_enabled(pvt)) { amd64_read_dct_pci_cfg(pvt, 1, DCLR0, &pvt->dclr1); amd64_read_dct_pci_cfg(pvt, 1, DCHR0, &pvt->dchr1); } determine_ecc_sym_sz(pvt); } /* * NOTE: CPU Revision Dependent code * * Input: * @csrow_nr ChipSelect Row Number (0..NUM_CHIPSELECTS-1) * k8 private pointer to --> * DRAM Bank Address mapping register * node_id * DCL register where dual_channel_active is * * The DBAM register consists of 4 sets of 4 bits each definitions: * * Bits: CSROWs * 0-3 CSROWs 0 and 1 * 4-7 CSROWs 2 and 3 * 8-11 CSROWs 4 and 5 * 12-15 CSROWs 6 and 7 * * Values range from: 0 to 15 * The meaning of the values depends on CPU revision and dual-channel state, * see relevant BKDG more info. * * The memory controller provides for total of only 8 CSROWs in its current * architecture. Each "pair" of CSROWs normally represents just one DIMM in * single channel or two (2) DIMMs in dual channel mode. * * The following code logic collapses the various tables for CSROW based on CPU * revision. * * Returns: * The number of PAGE_SIZE pages on the specified CSROW number it * encompasses * */ static u32 dct_get_csrow_nr_pages(struct amd64_pvt *pvt, u8 dct, int csrow_nr) { u32 dbam = dct ? pvt->dbam1 : pvt->dbam0; u32 cs_mode, nr_pages; csrow_nr >>= 1; cs_mode = DBAM_DIMM(csrow_nr, dbam); nr_pages = pvt->ops->dbam_to_cs(pvt, dct, cs_mode, csrow_nr); nr_pages <<= 20 - PAGE_SHIFT; edac_dbg(0, "csrow: %d, channel: %d, DBAM idx: %d\n", csrow_nr, dct, cs_mode); edac_dbg(0, "nr_pages/channel: %u\n", nr_pages); return nr_pages; } static u32 umc_get_csrow_nr_pages(struct amd64_pvt *pvt, u8 dct, int csrow_nr_orig) { int csrow_nr = csrow_nr_orig; u32 cs_mode, nr_pages; cs_mode = umc_get_cs_mode(csrow_nr >> 1, dct, pvt); nr_pages = umc_addr_mask_to_cs_size(pvt, dct, cs_mode, csrow_nr); nr_pages <<= 20 - PAGE_SHIFT; edac_dbg(0, "csrow: %d, channel: %d, cs_mode %d\n", csrow_nr_orig, dct, cs_mode); edac_dbg(0, "nr_pages/channel: %u\n", nr_pages); return nr_pages; } static void umc_init_csrows(struct mem_ctl_info *mci) { struct amd64_pvt *pvt = mci->pvt_info; enum edac_type edac_mode = EDAC_NONE; enum dev_type dev_type = DEV_UNKNOWN; struct dimm_info *dimm; u8 umc, cs; if (mci->edac_ctl_cap & EDAC_FLAG_S16ECD16ED) { edac_mode = EDAC_S16ECD16ED; dev_type = DEV_X16; } else if (mci->edac_ctl_cap & EDAC_FLAG_S8ECD8ED) { edac_mode = EDAC_S8ECD8ED; dev_type = DEV_X8; } else if (mci->edac_ctl_cap & EDAC_FLAG_S4ECD4ED) { edac_mode = EDAC_S4ECD4ED; dev_type = DEV_X4; } else if (mci->edac_ctl_cap & EDAC_FLAG_SECDED) { edac_mode = EDAC_SECDED; } for_each_umc(umc) { for_each_chip_select(cs, umc, pvt) { if (!csrow_enabled(cs, umc, pvt)) continue; dimm = mci->csrows[cs]->channels[umc]->dimm; edac_dbg(1, "MC node: %d, csrow: %d\n", pvt->mc_node_id, cs); dimm->nr_pages = umc_get_csrow_nr_pages(pvt, umc, cs); dimm->mtype = pvt->umc[umc].dram_type; dimm->edac_mode = edac_mode; dimm->dtype = dev_type; dimm->grain = 64; } } } /* * Initialize the array of csrow attribute instances, based on the values * from pci config hardware registers. */ static void dct_init_csrows(struct mem_ctl_info *mci) { struct amd64_pvt *pvt = mci->pvt_info; enum edac_type edac_mode = EDAC_NONE; struct csrow_info *csrow; struct dimm_info *dimm; int nr_pages = 0; int i, j; u32 val; amd64_read_pci_cfg(pvt->F3, NBCFG, &val); pvt->nbcfg = val; edac_dbg(0, "node %d, NBCFG=0x%08x[ChipKillEccCap: %d|DramEccEn: %d]\n", pvt->mc_node_id, val, !!(val & NBCFG_CHIPKILL), !!(val & NBCFG_ECC_ENABLE)); /* * We iterate over DCT0 here but we look at DCT1 in parallel, if needed. */ for_each_chip_select(i, 0, pvt) { bool row_dct0 = !!csrow_enabled(i, 0, pvt); bool row_dct1 = false; if (pvt->fam != 0xf) row_dct1 = !!csrow_enabled(i, 1, pvt); if (!row_dct0 && !row_dct1) continue; csrow = mci->csrows[i]; edac_dbg(1, "MC node: %d, csrow: %d\n", pvt->mc_node_id, i); if (row_dct0) { nr_pages = dct_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 = dct_get_csrow_nr_pages(pvt, 1, i); csrow->channels[1]->dimm->nr_pages = row_dct1_pages; nr_pages += row_dct1_pages; } edac_dbg(1, "Total csrow%d pages: %u\n", i, nr_pages); /* Determine DIMM ECC mode: */ if (pvt->nbcfg & NBCFG_ECC_ENABLE) { edac_mode = (pvt->nbcfg & NBCFG_CHIPKILL) ? EDAC_S4ECD4ED : EDAC_SECDED; } for (j = 0; j < pvt->max_mcs; j++) { dimm = csrow->channels[j]->dimm; dimm->mtype = pvt->dram_type; dimm->edac_mode = edac_mode; dimm->grain = 64; } } } /* get all cores on this DCT */ static void get_cpus_on_this_dct_cpumask(struct cpumask *mask, u16 nid) { int cpu; for_each_online_cpu(cpu) if (topology_die_id(cpu) == nid) cpumask_set_cpu(cpu, mask); } /* check MCG_CTL on all the cpus on this node */ static bool nb_mce_bank_enabled_on_node(u16 nid) { cpumask_var_t mask; int cpu, nbe; bool ret = false; if (!zalloc_cpumask_var(&mask, GFP_KERNEL)) { amd64_warn("%s: Error allocating mask\n", __func__); return false; } get_cpus_on_this_dct_cpumask(mask, nid); rdmsr_on_cpus(mask, MSR_IA32_MCG_CTL, msrs); for_each_cpu(cpu, mask) { struct msr *reg = per_cpu_ptr(msrs, cpu); nbe = reg->l & MSR_MCGCTL_NBE; edac_dbg(0, "core: %u, MCG_CTL: 0x%llx, NB MSR is %s\n", cpu, reg->q, (nbe ? "enabled" : "disabled")); if (!nbe) goto out; } ret = true; out: free_cpumask_var(mask); return ret; } static int toggle_ecc_err_reporting(struct ecc_settings *s, u16 nid, bool on) { cpumask_var_t cmask; int cpu; if (!zalloc_cpumask_var(&cmask, GFP_KERNEL)) { amd64_warn("%s: error allocating mask\n", __func__); return -ENOMEM; } get_cpus_on_this_dct_cpumask(cmask, nid); rdmsr_on_cpus(cmask, MSR_IA32_MCG_CTL, msrs); for_each_cpu(cpu, cmask) { struct msr *reg = per_cpu_ptr(msrs, cpu); if (on) { if (reg->l & MSR_MCGCTL_NBE) s->flags.nb_mce_enable = 1; reg->l |= MSR_MCGCTL_NBE; } else { /* * Turn off NB MCE reporting only when it was off before */ if (!s->flags.nb_mce_enable) reg->l &= ~MSR_MCGCTL_NBE; } } wrmsr_on_cpus(cmask, MSR_IA32_MCG_CTL, msrs); free_cpumask_var(cmask); return 0; } static bool enable_ecc_error_reporting(struct ecc_settings *s, u16 nid, struct pci_dev *F3) { bool ret = true; u32 value, mask = 0x3; /* UECC/CECC enable */ if (toggle_ecc_err_reporting(s, nid, ON)) { amd64_warn("Error enabling ECC reporting over MCGCTL!\n"); return false; } amd64_read_pci_cfg(F3, NBCTL, &value); s->old_nbctl = value & mask; s->nbctl_valid = true; value |= mask; amd64_write_pci_cfg(F3, NBCTL, value); amd64_read_pci_cfg(F3, NBCFG, &value); edac_dbg(0, "1: node %d, NBCFG=0x%08x[DramEccEn: %d]\n", nid, value, !!(value & NBCFG_ECC_ENABLE)); if (!(value & NBCFG_ECC_ENABLE)) { amd64_warn("DRAM ECC disabled on this node, enabling...\n"); s->flags.nb_ecc_prev = 0; /* Attempt to turn on DRAM ECC Enable */ value |= NBCFG_ECC_ENABLE; amd64_write_pci_cfg(F3, NBCFG, value); amd64_read_pci_cfg(F3, NBCFG, &value); if (!(value & NBCFG_ECC_ENABLE)) { amd64_warn("Hardware rejected DRAM ECC enable," "check memory DIMM configuration.\n"); ret = false; } else { amd64_info("Hardware accepted DRAM ECC Enable\n"); } } else { s->flags.nb_ecc_prev = 1; } edac_dbg(0, "2: node %d, NBCFG=0x%08x[DramEccEn: %d]\n", nid, value, !!(value & NBCFG_ECC_ENABLE)); return ret; } static void restore_ecc_error_reporting(struct ecc_settings *s, u16 nid, struct pci_dev *F3) { u32 value, mask = 0x3; /* UECC/CECC enable */ if (!s->nbctl_valid) return; amd64_read_pci_cfg(F3, NBCTL, &value); value &= ~mask; value |= s->old_nbctl; amd64_write_pci_cfg(F3, NBCTL, value); /* restore previous BIOS DRAM ECC "off" setting we force-enabled */ if (!s->flags.nb_ecc_prev) { amd64_read_pci_cfg(F3, NBCFG, &value); value &= ~NBCFG_ECC_ENABLE; amd64_write_pci_cfg(F3, NBCFG, value); } /* restore the NB Enable MCGCTL bit */ if (toggle_ecc_err_reporting(s, nid, OFF)) amd64_warn("Error restoring NB MCGCTL settings!\n"); } static bool dct_ecc_enabled(struct amd64_pvt *pvt) { u16 nid = pvt->mc_node_id; bool nb_mce_en = false; u8 ecc_en = 0; u32 value; amd64_read_pci_cfg(pvt->F3, NBCFG, &value); ecc_en = !!(value & NBCFG_ECC_ENABLE); nb_mce_en = nb_mce_bank_enabled_on_node(nid); if (!nb_mce_en) edac_dbg(0, "NB MCE bank disabled, set MSR 0x%08x[4] on node %d to enable.\n", MSR_IA32_MCG_CTL, nid); edac_dbg(3, "Node %d: DRAM ECC %s.\n", nid, (ecc_en ? "enabled" : "disabled")); if (!ecc_en || !nb_mce_en) return false; else return true; } static bool umc_ecc_enabled(struct amd64_pvt *pvt) { u8 umc_en_mask = 0, ecc_en_mask = 0; u16 nid = pvt->mc_node_id; struct amd64_umc *umc; u8 ecc_en = 0, i; for_each_umc(i) { umc = &pvt->umc[i]; /* Only check enabled UMCs. */ if (!(umc->sdp_ctrl & UMC_SDP_INIT)) continue; umc_en_mask |= BIT(i); if (umc->umc_cap_hi & UMC_ECC_ENABLED) ecc_en_mask |= BIT(i); } /* Check whether at least one UMC is enabled: */ if (umc_en_mask) ecc_en = umc_en_mask == ecc_en_mask; else edac_dbg(0, "Node %d: No enabled UMCs.\n", nid); edac_dbg(3, "Node %d: DRAM ECC %s.\n", nid, (ecc_en ? "enabled" : "disabled")); if (!ecc_en) return false; else return true; } static inline void umc_determine_edac_ctl_cap(struct mem_ctl_info *mci, struct amd64_pvt *pvt) { u8 i, ecc_en = 1, cpk_en = 1, dev_x4 = 1, dev_x16 = 1; for_each_umc(i) { if (pvt->umc[i].sdp_ctrl & UMC_SDP_INIT) { ecc_en &= !!(pvt->umc[i].umc_cap_hi & UMC_ECC_ENABLED); cpk_en &= !!(pvt->umc[i].umc_cap_hi & UMC_ECC_CHIPKILL_CAP); dev_x4 &= !!(pvt->umc[i].dimm_cfg & BIT(6)); dev_x16 &= !!(pvt->umc[i].dimm_cfg & BIT(7)); } } /* Set chipkill only if ECC is enabled: */ if (ecc_en) { mci->edac_ctl_cap |= EDAC_FLAG_SECDED; if (!cpk_en) return; if (dev_x4) mci->edac_ctl_cap |= EDAC_FLAG_S4ECD4ED; else if (dev_x16) mci->edac_ctl_cap |= EDAC_FLAG_S16ECD16ED; else mci->edac_ctl_cap |= EDAC_FLAG_S8ECD8ED; } } static void dct_setup_mci_misc_attrs(struct mem_ctl_info *mci) { struct amd64_pvt *pvt = mci->pvt_info; mci->mtype_cap = MEM_FLAG_DDR2 | MEM_FLAG_RDDR2; mci->edac_ctl_cap = EDAC_FLAG_NONE; if (pvt->nbcap & NBCAP_SECDED) mci->edac_ctl_cap |= EDAC_FLAG_SECDED; if (pvt->nbcap & NBCAP_CHIPKILL) mci->edac_ctl_cap |= EDAC_FLAG_S4ECD4ED; mci->edac_cap = dct_determine_edac_cap(pvt); mci->mod_name = EDAC_MOD_STR; mci->ctl_name = pvt->ctl_name; mci->dev_name = pci_name(pvt->F3); mci->ctl_page_to_phys = NULL; /* memory scrubber interface */ mci->set_sdram_scrub_rate = set_scrub_rate; mci->get_sdram_scrub_rate = get_scrub_rate; dct_init_csrows(mci); } static void umc_setup_mci_misc_attrs(struct mem_ctl_info *mci) { struct amd64_pvt *pvt = mci->pvt_info; mci->mtype_cap = MEM_FLAG_DDR4 | MEM_FLAG_RDDR4; mci->edac_ctl_cap = EDAC_FLAG_NONE; umc_determine_edac_ctl_cap(mci, pvt); mci->edac_cap = umc_determine_edac_cap(pvt); mci->mod_name = EDAC_MOD_STR; mci->ctl_name = pvt->ctl_name; mci->dev_name = pci_name(pvt->F3); mci->ctl_page_to_phys = NULL; umc_init_csrows(mci); } static int dct_hw_info_get(struct amd64_pvt *pvt) { int ret = reserve_mc_sibling_devs(pvt, pvt->f1_id, pvt->f2_id); if (ret) return ret; dct_prep_chip_selects(pvt); dct_read_base_mask(pvt); dct_read_mc_regs(pvt); dct_determine_memory_type(pvt); return 0; } static int umc_hw_info_get(struct amd64_pvt *pvt) { pvt->umc = kcalloc(pvt->max_mcs, sizeof(struct amd64_umc), GFP_KERNEL); if (!pvt->umc) return -ENOMEM; umc_prep_chip_selects(pvt); umc_read_base_mask(pvt); umc_read_mc_regs(pvt); umc_determine_memory_type(pvt); return 0; } /* * The CPUs have one channel per UMC, so UMC number is equivalent to a * channel number. The GPUs have 8 channels per UMC, so the UMC number no * longer works as a channel number. * * The channel number within a GPU UMC is given in MCA_IPID[15:12]. * However, the IDs are split such that two UMC values go to one UMC, and * the channel numbers are split in two groups of four. * * Refer to comment on gpu_get_umc_base(). * * For example, * UMC0 CH[3:0] = 0x0005[3:0]000 * UMC0 CH[7:4] = 0x0015[3:0]000 * UMC1 CH[3:0] = 0x0025[3:0]000 * UMC1 CH[7:4] = 0x0035[3:0]000 */ static void gpu_get_err_info(struct mce *m, struct err_info *err) { u8 ch = (m->ipid & GENMASK(31, 0)) >> 20; u8 phy = ((m->ipid >> 12) & 0xf); err->channel = ch % 2 ? phy + 4 : phy; err->csrow = phy; } static int gpu_addr_mask_to_cs_size(struct amd64_pvt *pvt, u8 umc, unsigned int cs_mode, int csrow_nr) { u32 addr_mask_orig = pvt->csels[umc].csmasks[csrow_nr]; return __addr_mask_to_cs_size(addr_mask_orig, cs_mode, csrow_nr, csrow_nr >> 1); } static void gpu_debug_display_dimm_sizes(struct amd64_pvt *pvt, u8 ctrl) { int size, cs_mode, cs = 0; edac_printk(KERN_DEBUG, EDAC_MC, "UMC%d chip selects:\n", ctrl); cs_mode = CS_EVEN_PRIMARY | CS_ODD_PRIMARY; for_each_chip_select(cs, ctrl, pvt) { size = gpu_addr_mask_to_cs_size(pvt, ctrl, cs_mode, cs); amd64_info(EDAC_MC ": %d: %5dMB\n", cs, size); } } static void gpu_dump_misc_regs(struct amd64_pvt *pvt) { struct amd64_umc *umc; u32 i; for_each_umc(i) { umc = &pvt->umc[i]; edac_dbg(1, "UMC%d UMC cfg: 0x%x\n", i, umc->umc_cfg); edac_dbg(1, "UMC%d SDP ctrl: 0x%x\n", i, umc->sdp_ctrl); edac_dbg(1, "UMC%d ECC ctrl: 0x%x\n", i, umc->ecc_ctrl); edac_dbg(1, "UMC%d All HBMs support ECC: yes\n", i); gpu_debug_display_dimm_sizes(pvt, i); } } static u32 gpu_get_csrow_nr_pages(struct amd64_pvt *pvt, u8 dct, int csrow_nr) { u32 nr_pages; int cs_mode = CS_EVEN_PRIMARY | CS_ODD_PRIMARY; nr_pages = gpu_addr_mask_to_cs_size(pvt, dct, cs_mode, csrow_nr); nr_pages <<= 20 - PAGE_SHIFT; edac_dbg(0, "csrow: %d, channel: %d\n", csrow_nr, dct); edac_dbg(0, "nr_pages/channel: %u\n", nr_pages); return nr_pages; } static void gpu_init_csrows(struct mem_ctl_info *mci) { struct amd64_pvt *pvt = mci->pvt_info; struct dimm_info *dimm; u8 umc, cs; for_each_umc(umc) { for_each_chip_select(cs, umc, pvt) { if (!csrow_enabled(cs, umc, pvt)) continue; dimm = mci->csrows[umc]->channels[cs]->dimm; edac_dbg(1, "MC node: %d, csrow: %d\n", pvt->mc_node_id, cs); dimm->nr_pages = gpu_get_csrow_nr_pages(pvt, umc, cs); dimm->edac_mode = EDAC_SECDED; dimm->mtype = MEM_HBM2; dimm->dtype = DEV_X16; dimm->grain = 64; } } } static void gpu_setup_mci_misc_attrs(struct mem_ctl_info *mci) { struct amd64_pvt *pvt = mci->pvt_info; mci->mtype_cap = MEM_FLAG_HBM2; mci->edac_ctl_cap = EDAC_FLAG_SECDED; mci->edac_cap = EDAC_FLAG_EC; mci->mod_name = EDAC_MOD_STR; mci->ctl_name = pvt->ctl_name; mci->dev_name = pci_name(pvt->F3); mci->ctl_page_to_phys = NULL; gpu_init_csrows(mci); } /* ECC is enabled by default on GPU nodes */ static bool gpu_ecc_enabled(struct amd64_pvt *pvt) { return true; } static inline u32 gpu_get_umc_base(u8 umc, u8 channel) { /* * On CPUs, there is one channel per UMC, so UMC numbering equals * channel numbering. On GPUs, there are eight channels per UMC, * so the channel numbering is different from UMC numbering. * * On CPU nodes channels are selected in 6th nibble * UMC chY[3:0]= [(chY*2 + 1) : (chY*2)]50000; * * On GPU nodes channels are selected in 3rd nibble * HBM chX[3:0]= [Y ]5X[3:0]000; * HBM chX[7:4]= [Y+1]5X[3:0]000 */ umc *= 2; if (channel >= 4) umc++; return 0x50000 + (umc << 20) + ((channel % 4) << 12); } static void gpu_read_mc_regs(struct amd64_pvt *pvt) { u8 nid = pvt->mc_node_id; struct amd64_umc *umc; u32 i, umc_base; /* Read registers from each UMC */ for_each_umc(i) { umc_base = gpu_get_umc_base(i, 0); umc = &pvt->umc[i]; amd_smn_read(nid, umc_base + UMCCH_UMC_CFG, &umc->umc_cfg); amd_smn_read(nid, umc_base + UMCCH_SDP_CTRL, &umc->sdp_ctrl); amd_smn_read(nid, umc_base + UMCCH_ECC_CTRL, &umc->ecc_ctrl); } } static void gpu_read_base_mask(struct amd64_pvt *pvt) { u32 base_reg, mask_reg; u32 *base, *mask; int umc, cs; for_each_umc(umc) { for_each_chip_select(cs, umc, pvt) { base_reg = gpu_get_umc_base(umc, cs) + UMCCH_BASE_ADDR; base = &pvt->csels[umc].csbases[cs]; if (!amd_smn_read(pvt->mc_node_id, base_reg, base)) { edac_dbg(0, " DCSB%d[%d]=0x%08x reg: 0x%x\n", umc, cs, *base, base_reg); } mask_reg = gpu_get_umc_base(umc, cs) + UMCCH_ADDR_MASK; mask = &pvt->csels[umc].csmasks[cs]; if (!amd_smn_read(pvt->mc_node_id, mask_reg, mask)) { edac_dbg(0, " DCSM%d[%d]=0x%08x reg: 0x%x\n", umc, cs, *mask, mask_reg); } } } } static void gpu_prep_chip_selects(struct amd64_pvt *pvt) { int umc; for_each_umc(umc) { pvt->csels[umc].b_cnt = 8; pvt->csels[umc].m_cnt = 8; } } static int gpu_hw_info_get(struct amd64_pvt *pvt) { int ret; ret = gpu_get_node_map(); if (ret) return ret; pvt->umc = kcalloc(pvt->max_mcs, sizeof(struct amd64_umc), GFP_KERNEL); if (!pvt->umc) return -ENOMEM; gpu_prep_chip_selects(pvt); gpu_read_base_mask(pvt); gpu_read_mc_regs(pvt); return 0; } static void hw_info_put(struct amd64_pvt *pvt) { pci_dev_put(pvt->F1); pci_dev_put(pvt->F2); kfree(pvt->umc); } static struct low_ops umc_ops = { .hw_info_get = umc_hw_info_get, .ecc_enabled = umc_ecc_enabled, .setup_mci_misc_attrs = umc_setup_mci_misc_attrs, .dump_misc_regs = umc_dump_misc_regs, .get_err_info = umc_get_err_info, }; static struct low_ops gpu_ops = { .hw_info_get = gpu_hw_info_get, .ecc_enabled = gpu_ecc_enabled, .setup_mci_misc_attrs = gpu_setup_mci_misc_attrs, .dump_misc_regs = gpu_dump_misc_regs, .get_err_info = gpu_get_err_info, }; /* Use Family 16h versions for defaults and adjust as needed below. */ static struct low_ops dct_ops = { .map_sysaddr_to_csrow = f1x_map_sysaddr_to_csrow, .dbam_to_cs = f16_dbam_to_chip_select, .hw_info_get = dct_hw_info_get, .ecc_enabled = dct_ecc_enabled, .setup_mci_misc_attrs = dct_setup_mci_misc_attrs, .dump_misc_regs = dct_dump_misc_regs, }; static int per_family_init(struct amd64_pvt *pvt) { pvt->ext_model = boot_cpu_data.x86_model >> 4; pvt->stepping = boot_cpu_data.x86_stepping; pvt->model = boot_cpu_data.x86_model; pvt->fam = boot_cpu_data.x86; pvt->max_mcs = 2; /* * Decide on which ops group to use here and do any family/model * overrides below. */ if (pvt->fam >= 0x17) pvt->ops = &umc_ops; else pvt->ops = &dct_ops; switch (pvt->fam) { case 0xf: pvt->ctl_name = (pvt->ext_model >= K8_REV_F) ? "K8 revF or later" : "K8 revE or earlier"; pvt->f1_id = PCI_DEVICE_ID_AMD_K8_NB_ADDRMAP; pvt->f2_id = PCI_DEVICE_ID_AMD_K8_NB_MEMCTL; pvt->ops->map_sysaddr_to_csrow = k8_map_sysaddr_to_csrow; pvt->ops->dbam_to_cs = k8_dbam_to_chip_select; break; case 0x10: pvt->ctl_name = "F10h"; pvt->f1_id = PCI_DEVICE_ID_AMD_10H_NB_MAP; pvt->f2_id = PCI_DEVICE_ID_AMD_10H_NB_DRAM; pvt->ops->dbam_to_cs = f10_dbam_to_chip_select; break; case 0x15: switch (pvt->model) { case 0x30: pvt->ctl_name = "F15h_M30h"; pvt->f1_id = PCI_DEVICE_ID_AMD_15H_M30H_NB_F1; pvt->f2_id = PCI_DEVICE_ID_AMD_15H_M30H_NB_F2; break; case 0x60: pvt->ctl_name = "F15h_M60h"; pvt->f1_id = PCI_DEVICE_ID_AMD_15H_M60H_NB_F1; pvt->f2_id = PCI_DEVICE_ID_AMD_15H_M60H_NB_F2; pvt->ops->dbam_to_cs = f15_m60h_dbam_to_chip_select; break; case 0x13: /* Richland is only client */ return -ENODEV; default: pvt->ctl_name = "F15h"; pvt->f1_id = PCI_DEVICE_ID_AMD_15H_NB_F1; pvt->f2_id = PCI_DEVICE_ID_AMD_15H_NB_F2; pvt->ops->dbam_to_cs = f15_dbam_to_chip_select; break; } break; case 0x16: switch (pvt->model) { case 0x30: pvt->ctl_name = "F16h_M30h"; pvt->f1_id = PCI_DEVICE_ID_AMD_16H_M30H_NB_F1; pvt->f2_id = PCI_DEVICE_ID_AMD_16H_M30H_NB_F2; break; default: pvt->ctl_name = "F16h"; pvt->f1_id = PCI_DEVICE_ID_AMD_16H_NB_F1; pvt->f2_id = PCI_DEVICE_ID_AMD_16H_NB_F2; break; } break; case 0x17: switch (pvt->model) { case 0x10 ... 0x2f: pvt->ctl_name = "F17h_M10h"; break; case 0x30 ... 0x3f: pvt->ctl_name = "F17h_M30h"; pvt->max_mcs = 8; break; case 0x60 ... 0x6f: pvt->ctl_name = "F17h_M60h"; break; case 0x70 ... 0x7f: pvt->ctl_name = "F17h_M70h"; break; default: pvt->ctl_name = "F17h"; break; } break; case 0x18: pvt->ctl_name = "F18h"; break; case 0x19: switch (pvt->model) { case 0x00 ... 0x0f: pvt->ctl_name = "F19h"; pvt->max_mcs = 8; break; case 0x10 ... 0x1f: pvt->ctl_name = "F19h_M10h"; pvt->max_mcs = 12; pvt->flags.zn_regs_v2 = 1; break; case 0x20 ... 0x2f: pvt->ctl_name = "F19h_M20h"; break; case 0x30 ... 0x3f: if (pvt->F3->device == PCI_DEVICE_ID_AMD_MI200_DF_F3) { pvt->ctl_name = "MI200"; pvt->max_mcs = 4; pvt->ops = &gpu_ops; } else { pvt->ctl_name = "F19h_M30h"; pvt->max_mcs = 8; } break; case 0x50 ... 0x5f: pvt->ctl_name = "F19h_M50h"; break; case 0x60 ... 0x6f: pvt->ctl_name = "F19h_M60h"; pvt->flags.zn_regs_v2 = 1; break; case 0x70 ... 0x7f: pvt->ctl_name = "F19h_M70h"; pvt->flags.zn_regs_v2 = 1; break; case 0xa0 ... 0xaf: pvt->ctl_name = "F19h_MA0h"; pvt->max_mcs = 12; pvt->flags.zn_regs_v2 = 1; break; } break; default: amd64_err("Unsupported family!\n"); return -ENODEV; } return 0; } static const struct attribute_group *amd64_edac_attr_groups[] = { #ifdef CONFIG_EDAC_DEBUG &dbg_group, &inj_group, #endif NULL }; static int init_one_instance(struct amd64_pvt *pvt) { struct mem_ctl_info *mci = NULL; struct edac_mc_layer layers[2]; int ret = -ENOMEM; /* * For Heterogeneous family EDAC CHIP_SELECT and CHANNEL layers should * be swapped to fit into the layers. */ layers[0].type = EDAC_MC_LAYER_CHIP_SELECT; layers[0].size = (pvt->F3->device == PCI_DEVICE_ID_AMD_MI200_DF_F3) ? pvt->max_mcs : pvt->csels[0].b_cnt; layers[0].is_virt_csrow = true; layers[1].type = EDAC_MC_LAYER_CHANNEL; layers[1].size = (pvt->F3->device == PCI_DEVICE_ID_AMD_MI200_DF_F3) ? pvt->csels[0].b_cnt : pvt->max_mcs; layers[1].is_virt_csrow = false; mci = edac_mc_alloc(pvt->mc_node_id, ARRAY_SIZE(layers), layers, 0); if (!mci) return ret; mci->pvt_info = pvt; mci->pdev = &pvt->F3->dev; pvt->ops->setup_mci_misc_attrs(mci); ret = -ENODEV; if (edac_mc_add_mc_with_groups(mci, amd64_edac_attr_groups)) { edac_dbg(1, "failed edac_mc_add_mc()\n"); edac_mc_free(mci); return ret; } return 0; } static bool instance_has_memory(struct amd64_pvt *pvt) { bool cs_enabled = false; int cs = 0, dct = 0; for (dct = 0; dct < pvt->max_mcs; dct++) { for_each_chip_select(cs, dct, pvt) cs_enabled |= csrow_enabled(cs, dct, pvt); } return cs_enabled; } static int probe_one_instance(unsigned int nid) { struct pci_dev *F3 = node_to_amd_nb(nid)->misc; struct amd64_pvt *pvt = NULL; struct ecc_settings *s; int ret; ret = -ENOMEM; s = kzalloc(sizeof(struct ecc_settings), GFP_KERNEL); if (!s) goto err_out; ecc_stngs[nid] = s; pvt = kzalloc(sizeof(struct amd64_pvt), GFP_KERNEL); if (!pvt) goto err_settings; pvt->mc_node_id = nid; pvt->F3 = F3; ret = per_family_init(pvt); if (ret < 0) goto err_enable; ret = pvt->ops->hw_info_get(pvt); if (ret < 0) goto err_enable; ret = 0; if (!instance_has_memory(pvt)) { amd64_info("Node %d: No DIMMs detected.\n", nid); goto err_enable; } if (!pvt->ops->ecc_enabled(pvt)) { ret = -ENODEV; if (!ecc_enable_override) goto err_enable; if (boot_cpu_data.x86 >= 0x17) { amd64_warn("Forcing ECC on is not recommended on newer systems. Please enable ECC in BIOS."); goto err_enable; } else amd64_warn("Forcing ECC on!\n"); if (!enable_ecc_error_reporting(s, nid, F3)) goto err_enable; } ret = init_one_instance(pvt); if (ret < 0) { amd64_err("Error probing instance: %d\n", nid); if (boot_cpu_data.x86 < 0x17) restore_ecc_error_reporting(s, nid, F3); goto err_enable; } amd64_info("%s detected (node %d).\n", pvt->ctl_name, pvt->mc_node_id); /* Display and decode various registers for debug purposes. */ pvt->ops->dump_misc_regs(pvt); return ret; err_enable: hw_info_put(pvt); kfree(pvt); err_settings: kfree(s); ecc_stngs[nid] = NULL; err_out: return ret; } static void remove_one_instance(unsigned int nid) { struct pci_dev *F3 = node_to_amd_nb(nid)->misc; struct ecc_settings *s = ecc_stngs[nid]; struct mem_ctl_info *mci; struct amd64_pvt *pvt; /* Remove from EDAC CORE tracking list */ mci = edac_mc_del_mc(&F3->dev); if (!mci) return; pvt = mci->pvt_info; restore_ecc_error_reporting(s, nid, F3); kfree(ecc_stngs[nid]); ecc_stngs[nid] = NULL; /* Free the EDAC CORE resources */ mci->pvt_info = NULL; hw_info_put(pvt); kfree(pvt); edac_mc_free(mci); } static void setup_pci_device(void) { if (pci_ctl) return; pci_ctl = edac_pci_create_generic_ctl(pci_ctl_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__); } } static const struct x86_cpu_id amd64_cpuids[] = { X86_MATCH_VENDOR_FAM(AMD, 0x0F, NULL), X86_MATCH_VENDOR_FAM(AMD, 0x10, NULL), X86_MATCH_VENDOR_FAM(AMD, 0x15, NULL), X86_MATCH_VENDOR_FAM(AMD, 0x16, NULL), X86_MATCH_VENDOR_FAM(AMD, 0x17, NULL), X86_MATCH_VENDOR_FAM(HYGON, 0x18, NULL), X86_MATCH_VENDOR_FAM(AMD, 0x19, NULL), { } }; MODULE_DEVICE_TABLE(x86cpu, amd64_cpuids); static int __init amd64_edac_init(void) { const char *owner; int err = -ENODEV; int i; if (ghes_get_devices()) return -EBUSY; owner = edac_get_owner(); if (owner && strncmp(owner, EDAC_MOD_STR, sizeof(EDAC_MOD_STR))) return -EBUSY; if (!x86_match_cpu(amd64_cpuids)) return -ENODEV; if (!amd_nb_num()) return -ENODEV; opstate_init(); err = -ENOMEM; ecc_stngs = kcalloc(amd_nb_num(), sizeof(ecc_stngs[0]), GFP_KERNEL); if (!ecc_stngs) goto err_free; msrs = msrs_alloc(); if (!msrs) goto err_free; for (i = 0; i < amd_nb_num(); i++) { err = probe_one_instance(i); if (err) { /* unwind properly */ while (--i >= 0) remove_one_instance(i); goto err_pci; } } if (!edac_has_mcs()) { err = -ENODEV; goto err_pci; } /* register stuff with EDAC MCE */ if (boot_cpu_data.x86 >= 0x17) { amd_register_ecc_decoder(decode_umc_error); } else { amd_register_ecc_decoder(decode_bus_error); setup_pci_device(); } #ifdef CONFIG_X86_32 amd64_err("%s on 32-bit is unsupported. USE AT YOUR OWN RISK!\n", EDAC_MOD_STR); #endif return 0; err_pci: pci_ctl_dev = NULL; msrs_free(msrs); msrs = NULL; err_free: kfree(ecc_stngs); ecc_stngs = NULL; return err; } static void __exit amd64_edac_exit(void) { int i; if (pci_ctl) edac_pci_release_generic_ctl(pci_ctl); /* unregister from EDAC MCE */ if (boot_cpu_data.x86 >= 0x17) amd_unregister_ecc_decoder(decode_umc_error); else amd_unregister_ecc_decoder(decode_bus_error); for (i = 0; i < amd_nb_num(); i++) remove_one_instance(i); kfree(ecc_stngs); ecc_stngs = NULL; pci_ctl_dev = NULL; msrs_free(msrs); msrs = NULL; } module_init(amd64_edac_init); module_exit(amd64_edac_exit); MODULE_LICENSE("GPL"); MODULE_AUTHOR("SoftwareBitMaker: Doug Thompson, Dave Peterson, Thayne Harbaugh; AMD"); MODULE_DESCRIPTION("MC support for AMD64 memory controllers"); module_param(edac_op_state, int, 0444); MODULE_PARM_DESC(edac_op_state, "EDAC Error Reporting state: 0=Poll,1=NMI");