diff options
Diffstat (limited to 'drivers/edac/i82875p_edac.c')
| -rw-r--r-- | drivers/edac/i82875p_edac.c | 364 |
1 files changed, 216 insertions, 148 deletions
diff --git a/drivers/edac/i82875p_edac.c b/drivers/edac/i82875p_edac.c index 0aec92698f1..64b68320249 100644 --- a/drivers/edac/i82875p_edac.c +++ b/drivers/edac/i82875p_edac.c @@ -13,13 +13,15 @@ * Note: E7210 appears same as D82875P - zhenyu.z.wang at intel.com */ -#include <linux/config.h> #include <linux/module.h> #include <linux/init.h> #include <linux/pci.h> #include <linux/pci_ids.h> -#include <linux/slab.h> -#include "edac_mc.h" +#include <linux/edac.h> +#include "edac_core.h" + +#define I82875P_REVISION " Ver: 2.0.2" +#define EDAC_MOD_STR "i82875p_edac" #define i82875p_printk(level, fmt, arg...) \ edac_printk(level, "i82875p", fmt, ##arg) @@ -36,7 +38,8 @@ #endif /* PCI_DEVICE_ID_INTEL_82875_6 */ /* four csrows in dual channel, eight in single channel */ -#define I82875P_NR_CSROWS(nr_chans) (8/(nr_chans)) +#define I82875P_NR_DIMMS 8 +#define I82875P_NR_CSROWS(nr_chans) (I82875P_NR_DIMMS / (nr_chans)) /* Intel 82875p register addresses - device 0 function 0 - DRAM Controller */ #define I82875P_EAP 0x58 /* Error Address Pointer (32b) @@ -172,31 +175,36 @@ struct i82875p_error_info { static const struct i82875p_dev_info i82875p_devs[] = { [I82875P] = { - .ctl_name = "i82875p" - }, + .ctl_name = "i82875p"}, }; -static struct pci_dev *mci_pdev = NULL; /* init dev: in case that AGP code has +static struct pci_dev *mci_pdev; /* init dev: in case that AGP code has * already registered driver */ -static int i82875p_registered = 1; +static struct edac_pci_ctl_info *i82875p_pci; static void i82875p_get_error_info(struct mem_ctl_info *mci, - struct i82875p_error_info *info) + struct i82875p_error_info *info) { + struct pci_dev *pdev; + + pdev = to_pci_dev(mci->pdev); + /* * This is a mess because there is no atomic way to read all the * registers at once and the registers can transition from CE being * overwritten by UE. */ - pci_read_config_word(mci->pdev, I82875P_ERRSTS, &info->errsts); - pci_read_config_dword(mci->pdev, I82875P_EAP, &info->eap); - pci_read_config_byte(mci->pdev, I82875P_DES, &info->des); - pci_read_config_byte(mci->pdev, I82875P_DERRSYN, &info->derrsyn); - pci_read_config_word(mci->pdev, I82875P_ERRSTS, &info->errsts2); + pci_read_config_word(pdev, I82875P_ERRSTS, &info->errsts); - pci_write_bits16(mci->pdev, I82875P_ERRSTS, 0x0081, 0x0081); + if (!(info->errsts & 0x0081)) + return; + + pci_read_config_dword(pdev, I82875P_EAP, &info->eap); + pci_read_config_byte(pdev, I82875P_DES, &info->des); + pci_read_config_byte(pdev, I82875P_DERRSYN, &info->derrsyn); + pci_read_config_word(pdev, I82875P_ERRSTS, &info->errsts2); /* * If the error is the same then we can for both reads then @@ -204,32 +212,33 @@ static void i82875p_get_error_info(struct mem_ctl_info *mci, * there is a CE no info and the second set of reads is valid * and should be UE info. */ - if (!(info->errsts2 & 0x0081)) - return; - if ((info->errsts ^ info->errsts2) & 0x0081) { - pci_read_config_dword(mci->pdev, I82875P_EAP, &info->eap); - pci_read_config_byte(mci->pdev, I82875P_DES, &info->des); - pci_read_config_byte(mci->pdev, I82875P_DERRSYN, - &info->derrsyn); + pci_read_config_dword(pdev, I82875P_EAP, &info->eap); + pci_read_config_byte(pdev, I82875P_DES, &info->des); + pci_read_config_byte(pdev, I82875P_DERRSYN, &info->derrsyn); } + + pci_write_bits16(pdev, I82875P_ERRSTS, 0x0081, 0x0081); } static int i82875p_process_error_info(struct mem_ctl_info *mci, - struct i82875p_error_info *info, int handle_errors) + struct i82875p_error_info *info, + int handle_errors) { int row, multi_chan; - multi_chan = mci->csrows[0].nr_channels - 1; + multi_chan = mci->csrows[0]->nr_channels - 1; - if (!(info->errsts2 & 0x0081)) + if (!(info->errsts & 0x0081)) return 0; if (!handle_errors) return 1; if ((info->errsts ^ info->errsts2) & 0x0081) { - edac_mc_handle_ce_no_info(mci, "UE overwrote CE"); + edac_mc_handle_error(HW_EVENT_ERR_UNCORRECTED, mci, 1, 0, 0, 0, + -1, -1, -1, + "UE overwrote CE", ""); info->errsts = info->errsts2; } @@ -237,11 +246,15 @@ static int i82875p_process_error_info(struct mem_ctl_info *mci, row = edac_mc_find_csrow_by_page(mci, info->eap); if (info->errsts & 0x0080) - edac_mc_handle_ue(mci, info->eap, 0, row, "i82875p UE"); + edac_mc_handle_error(HW_EVENT_ERR_UNCORRECTED, mci, 1, + info->eap, 0, 0, + row, -1, -1, + "i82875p UE", ""); else - edac_mc_handle_ce(mci, info->eap, 0, info->derrsyn, row, - multi_chan ? (info->des & 0x1) : 0, - "i82875p CE"); + edac_mc_handle_error(HW_EVENT_ERR_CORRECTED, mci, 1, + info->eap, 0, info->derrsyn, + row, multi_chan ? (info->des & 0x1) : 0, + -1, "i82875p CE", ""); return 1; } @@ -250,173 +263,220 @@ static void i82875p_check(struct mem_ctl_info *mci) { struct i82875p_error_info info; - debugf1("MC%d: %s()\n", mci->mc_idx, __func__); + edac_dbg(1, "MC%d\n", mci->mc_idx); i82875p_get_error_info(mci, &info); i82875p_process_error_info(mci, &info, 1); } -#ifdef CONFIG_PROC_FS -extern int pci_proc_attach_device(struct pci_dev *); -#endif - -static int i82875p_probe1(struct pci_dev *pdev, int dev_idx) +/* Return 0 on success or 1 on failure. */ +static int i82875p_setup_overfl_dev(struct pci_dev *pdev, + struct pci_dev **ovrfl_pdev, + void __iomem **ovrfl_window) { - int rc = -ENODEV; - int index; - struct mem_ctl_info *mci = NULL; - struct i82875p_pvt *pvt = NULL; - unsigned long last_cumul_size; - struct pci_dev *ovrfl_pdev; - void __iomem *ovrfl_window = NULL; - u32 drc; - u32 drc_chan; /* Number of channels 0=1chan,1=2chan */ - u32 nr_chans; - u32 drc_ddim; /* DRAM Data Integrity Mode 0=none,2=edac */ - struct i82875p_error_info discard; + struct pci_dev *dev; + void __iomem *window; - debugf0("%s()\n", __func__); - ovrfl_pdev = pci_get_device(PCI_VEND_DEV(INTEL, 82875_6), NULL); + *ovrfl_pdev = NULL; + *ovrfl_window = NULL; + dev = pci_get_device(PCI_VEND_DEV(INTEL, 82875_6), NULL); - if (!ovrfl_pdev) { - /* - * Intel tells BIOS developers to hide device 6 which + if (dev == NULL) { + /* Intel tells BIOS developers to hide device 6 which * configures the overflow device access containing * the DRBs - this is where we expose device 6. * http://www.x86-secret.com/articles/tweak/pat/patsecrets-2.htm */ pci_write_bits8(pdev, 0xf4, 0x2, 0x2); - ovrfl_pdev = - pci_scan_single_device(pdev->bus, PCI_DEVFN(6, 0)); + dev = pci_scan_single_device(pdev->bus, PCI_DEVFN(6, 0)); - if (!ovrfl_pdev) - return -ENODEV; - } + if (dev == NULL) + return 1; -#ifdef CONFIG_PROC_FS - if (!ovrfl_pdev->procent && pci_proc_attach_device(ovrfl_pdev)) { - i82875p_printk(KERN_ERR, - "%s(): Failed to attach overflow device\n", __func__); - return -ENODEV; + pci_bus_assign_resources(dev->bus); + pci_bus_add_device(dev); } -#endif - /* CONFIG_PROC_FS */ - if (pci_enable_device(ovrfl_pdev)) { - i82875p_printk(KERN_ERR, - "%s(): Failed to enable overflow device\n", __func__); - return -ENODEV; + + *ovrfl_pdev = dev; + + if (pci_enable_device(dev)) { + i82875p_printk(KERN_ERR, "%s(): Failed to enable overflow " + "device\n", __func__); + return 1; } - if (pci_request_regions(ovrfl_pdev, pci_name(ovrfl_pdev))) { + if (pci_request_regions(dev, pci_name(dev))) { #ifdef CORRECT_BIOS goto fail0; #endif } /* cache is irrelevant for PCI bus reads/writes */ - ovrfl_window = ioremap_nocache(pci_resource_start(ovrfl_pdev, 0), - pci_resource_len(ovrfl_pdev, 0)); - - if (!ovrfl_window) { + window = pci_ioremap_bar(dev, 0); + if (window == NULL) { i82875p_printk(KERN_ERR, "%s(): Failed to ioremap bar6\n", __func__); goto fail1; } - /* need to find out the number of channels */ - drc = readl(ovrfl_window + I82875P_DRC); - drc_chan = ((drc >> 21) & 0x1); - nr_chans = drc_chan + 1; + *ovrfl_window = window; + return 0; - drc_ddim = (drc >> 18) & 0x1; - mci = edac_mc_alloc(sizeof(*pvt), I82875P_NR_CSROWS(nr_chans), - nr_chans); +fail1: + pci_release_regions(dev); - if (!mci) { - rc = -ENOMEM; - goto fail2; - } +#ifdef CORRECT_BIOS +fail0: + pci_disable_device(dev); +#endif + /* NOTE: the ovrfl proc entry and pci_dev are intentionally left */ + return 1; +} - debugf3("%s(): init mci\n", __func__); - mci->pdev = pdev; - mci->mtype_cap = MEM_FLAG_DDR; - mci->edac_ctl_cap = EDAC_FLAG_NONE | EDAC_FLAG_SECDED; - mci->edac_cap = EDAC_FLAG_UNKNOWN; - /* adjust FLAGS */ +/* Return 1 if dual channel mode is active. Else return 0. */ +static inline int dual_channel_active(u32 drc) +{ + return (drc >> 21) & 0x1; +} - mci->mod_name = EDAC_MOD_STR; - mci->mod_ver = "$Revision: 1.5.2.11 $"; - mci->ctl_name = i82875p_devs[dev_idx].ctl_name; - mci->edac_check = i82875p_check; - mci->ctl_page_to_phys = NULL; - debugf3("%s(): init pvt\n", __func__); - pvt = (struct i82875p_pvt *) mci->pvt_info; - pvt->ovrfl_pdev = ovrfl_pdev; - pvt->ovrfl_window = ovrfl_window; +static void i82875p_init_csrows(struct mem_ctl_info *mci, + struct pci_dev *pdev, + void __iomem * ovrfl_window, u32 drc) +{ + struct csrow_info *csrow; + struct dimm_info *dimm; + unsigned nr_chans = dual_channel_active(drc) + 1; + unsigned long last_cumul_size; + u8 value; + u32 drc_ddim; /* DRAM Data Integrity Mode 0=none,2=edac */ + u32 cumul_size, nr_pages; + int index, j; - /* - * The dram row boundary (DRB) reg values are boundary address + drc_ddim = (drc >> 18) & 0x1; + last_cumul_size = 0; + + /* The dram row boundary (DRB) reg values are boundary address * for each DRAM row with a granularity of 32 or 64MB (single/dual * channel operation). DRB regs are cumulative; therefore DRB7 will * contain the total memory contained in all eight rows. */ - for (last_cumul_size = index = 0; index < mci->nr_csrows; index++) { - u8 value; - u32 cumul_size; - struct csrow_info *csrow = &mci->csrows[index]; + + for (index = 0; index < mci->nr_csrows; index++) { + csrow = mci->csrows[index]; value = readb(ovrfl_window + I82875P_DRB + index); cumul_size = value << (I82875P_DRB_SHIFT - PAGE_SHIFT); - debugf3("%s(): (%d) cumul_size 0x%x\n", __func__, index, - cumul_size); - + edac_dbg(3, "(%d) cumul_size 0x%x\n", index, cumul_size); if (cumul_size == last_cumul_size) continue; /* not populated */ csrow->first_page = last_cumul_size; csrow->last_page = cumul_size - 1; - csrow->nr_pages = cumul_size - last_cumul_size; + nr_pages = cumul_size - last_cumul_size; last_cumul_size = cumul_size; - csrow->grain = 1 << 12; /* I82875P_EAP has 4KiB reolution */ - csrow->mtype = MEM_DDR; - csrow->dtype = DEV_UNKNOWN; - csrow->edac_mode = drc_ddim ? EDAC_SECDED : EDAC_NONE; + + for (j = 0; j < nr_chans; j++) { + dimm = csrow->channels[j]->dimm; + + dimm->nr_pages = nr_pages / nr_chans; + dimm->grain = 1 << 12; /* I82875P_EAP has 4KiB reolution */ + dimm->mtype = MEM_DDR; + dimm->dtype = DEV_UNKNOWN; + dimm->edac_mode = drc_ddim ? EDAC_SECDED : EDAC_NONE; + } } +} + +static int i82875p_probe1(struct pci_dev *pdev, int dev_idx) +{ + int rc = -ENODEV; + struct mem_ctl_info *mci; + struct edac_mc_layer layers[2]; + struct i82875p_pvt *pvt; + struct pci_dev *ovrfl_pdev; + void __iomem *ovrfl_window; + u32 drc; + u32 nr_chans; + struct i82875p_error_info discard; - i82875p_get_error_info(mci, &discard); /* clear counters */ + edac_dbg(0, "\n"); + + if (i82875p_setup_overfl_dev(pdev, &ovrfl_pdev, &ovrfl_window)) + return -ENODEV; + drc = readl(ovrfl_window + I82875P_DRC); + nr_chans = dual_channel_active(drc) + 1; + + layers[0].type = EDAC_MC_LAYER_CHIP_SELECT; + layers[0].size = I82875P_NR_CSROWS(nr_chans); + layers[0].is_virt_csrow = true; + layers[1].type = EDAC_MC_LAYER_CHANNEL; + layers[1].size = nr_chans; + layers[1].is_virt_csrow = false; + mci = edac_mc_alloc(0, ARRAY_SIZE(layers), layers, sizeof(*pvt)); + if (!mci) { + rc = -ENOMEM; + goto fail0; + } + edac_dbg(3, "init mci\n"); + mci->pdev = &pdev->dev; + mci->mtype_cap = MEM_FLAG_DDR; + mci->edac_ctl_cap = EDAC_FLAG_NONE | EDAC_FLAG_SECDED; + mci->edac_cap = EDAC_FLAG_UNKNOWN; + mci->mod_name = EDAC_MOD_STR; + mci->mod_ver = I82875P_REVISION; + mci->ctl_name = i82875p_devs[dev_idx].ctl_name; + mci->dev_name = pci_name(pdev); + mci->edac_check = i82875p_check; + mci->ctl_page_to_phys = NULL; + edac_dbg(3, "init pvt\n"); + pvt = (struct i82875p_pvt *)mci->pvt_info; + pvt->ovrfl_pdev = ovrfl_pdev; + pvt->ovrfl_window = ovrfl_window; + i82875p_init_csrows(mci, pdev, ovrfl_window, drc); + i82875p_get_error_info(mci, &discard); /* clear counters */ + + /* Here we assume that we will never see multiple instances of this + * type of memory controller. The ID is therefore hardcoded to 0. + */ if (edac_mc_add_mc(mci)) { - debugf3("%s(): failed edac_mc_add_mc()\n", __func__); - goto fail3; + edac_dbg(3, "failed edac_mc_add_mc()\n"); + goto fail1; + } + + /* allocating generic PCI control info */ + i82875p_pci = edac_pci_create_generic_ctl(&pdev->dev, EDAC_MOD_STR); + if (!i82875p_pci) { + printk(KERN_WARNING + "%s(): Unable to create PCI control\n", + __func__); + printk(KERN_WARNING + "%s(): PCI error report via EDAC not setup\n", + __func__); } /* get this far and it's successful */ - debugf3("%s(): success\n", __func__); + edac_dbg(3, "success\n"); return 0; -fail3: +fail1: edac_mc_free(mci); -fail2: +fail0: iounmap(ovrfl_window); - -fail1: pci_release_regions(ovrfl_pdev); -#ifdef CORRECT_BIOS -fail0: -#endif pci_disable_device(ovrfl_pdev); /* NOTE: the ovrfl proc entry and pci_dev are intentionally left */ return rc; } /* returns count (>= 0), or negative on error */ -static int __devinit i82875p_init_one(struct pci_dev *pdev, - const struct pci_device_id *ent) +static int i82875p_init_one(struct pci_dev *pdev, + const struct pci_device_id *ent) { int rc; - debugf0("%s()\n", __func__); + edac_dbg(0, "\n"); i82875p_printk(KERN_INFO, "i82875p init one\n"); if (pci_enable_device(pdev) < 0) @@ -430,17 +490,20 @@ static int __devinit i82875p_init_one(struct pci_dev *pdev, return rc; } -static void __devexit i82875p_remove_one(struct pci_dev *pdev) +static void i82875p_remove_one(struct pci_dev *pdev) { struct mem_ctl_info *mci; struct i82875p_pvt *pvt = NULL; - debugf0("%s()\n", __func__); + edac_dbg(0, "\n"); + + if (i82875p_pci) + edac_pci_release_generic_ctl(i82875p_pci); - if ((mci = edac_mc_del_mc(pdev)) == NULL) + if ((mci = edac_mc_del_mc(&pdev->dev)) == NULL) return; - pvt = (struct i82875p_pvt *) mci->pvt_info; + pvt = (struct i82875p_pvt *)mci->pvt_info; if (pvt->ovrfl_window) iounmap(pvt->ovrfl_window); @@ -456,14 +519,13 @@ static void __devexit i82875p_remove_one(struct pci_dev *pdev) edac_mc_free(mci); } -static const struct pci_device_id i82875p_pci_tbl[] __devinitdata = { +static const struct pci_device_id i82875p_pci_tbl[] = { { - PCI_VEND_DEV(INTEL, 82875_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, - I82875P - }, + PCI_VEND_DEV(INTEL, 82875_0), PCI_ANY_ID, PCI_ANY_ID, 0, 0, + I82875P}, { - 0, - } /* 0 terminated list. */ + 0, + } /* 0 terminated list. */ }; MODULE_DEVICE_TABLE(pci, i82875p_pci_tbl); @@ -471,7 +533,7 @@ MODULE_DEVICE_TABLE(pci, i82875p_pci_tbl); static struct pci_driver i82875p_driver = { .name = EDAC_MOD_STR, .probe = i82875p_init_one, - .remove = __devexit_p(i82875p_remove_one), + .remove = i82875p_remove_one, .id_table = i82875p_pci_tbl, }; @@ -479,7 +541,11 @@ static int __init i82875p_init(void) { int pci_rc; - debugf3("%s()\n", __func__); + edac_dbg(3, "\n"); + + /* Ensure that the OPSTATE is set correctly for POLL or NMI */ + opstate_init(); + pci_rc = pci_register_driver(&i82875p_driver); if (pci_rc < 0) @@ -487,10 +553,10 @@ static int __init i82875p_init(void) if (mci_pdev == NULL) { mci_pdev = pci_get_device(PCI_VENDOR_ID_INTEL, - PCI_DEVICE_ID_INTEL_82875_0, NULL); + PCI_DEVICE_ID_INTEL_82875_0, NULL); if (!mci_pdev) { - debugf0("875p pci_get_device fail\n"); + edac_dbg(0, "875p pci_get_device fail\n"); pci_rc = -ENODEV; goto fail1; } @@ -498,7 +564,7 @@ static int __init i82875p_init(void) pci_rc = i82875p_init_one(mci_pdev, i82875p_pci_tbl); if (pci_rc < 0) { - debugf0("875p init fail\n"); + edac_dbg(0, "875p init fail\n"); pci_rc = -ENODEV; goto fail1; } @@ -518,14 +584,13 @@ fail0: static void __exit i82875p_exit(void) { - debugf3("%s()\n", __func__); + edac_dbg(3, "\n"); + + i82875p_remove_one(mci_pdev); + pci_dev_put(mci_pdev); pci_unregister_driver(&i82875p_driver); - if (!i82875p_registered) { - i82875p_remove_one(mci_pdev); - pci_dev_put(mci_pdev); - } } module_init(i82875p_init); @@ -534,3 +599,6 @@ module_exit(i82875p_exit); MODULE_LICENSE("GPL"); MODULE_AUTHOR("Linux Networx (http://lnxi.com) Thayne Harbaugh"); MODULE_DESCRIPTION("MC support for Intel 82875 memory hub controllers"); + +module_param(edac_op_state, int, 0444); +MODULE_PARM_DESC(edac_op_state, "EDAC Error Reporting state: 0=Poll,1=NMI"); |
