linux_dsm_epyc7002/drivers/edac/i7core_edac.c

2466 lines
63 KiB
C
Raw Normal View History

/* Intel i7 core/Nehalem Memory Controller kernel module
*
* This driver supports the memory controllers found on the Intel
* processor families i7core, i7core 7xx/8xx, i5core, Xeon 35xx,
* Xeon 55xx and Xeon 56xx also known as Nehalem, Nehalem-EP, Lynnfield
* and Westmere-EP.
*
* This file may be distributed under the terms of the
* GNU General Public License version 2 only.
*
* Copyright (c) 2009-2010 by:
* Mauro Carvalho Chehab <mchehab@redhat.com>
*
* Red Hat Inc. http://www.redhat.com
*
* Forked and adapted from the i5400_edac driver
*
* Based on the following public Intel datasheets:
* Intel Core i7 Processor Extreme Edition and Intel Core i7 Processor
* Datasheet, Volume 2:
* http://download.intel.com/design/processor/datashts/320835.pdf
* Intel Xeon Processor 5500 Series Datasheet Volume 2
* http://www.intel.com/Assets/PDF/datasheet/321322.pdf
* also available at:
* http://www.arrownac.com/manufacturers/intel/s/nehalem/5500-datasheet-v2.pdf
*/
#include <linux/module.h>
#include <linux/init.h>
#include <linux/pci.h>
#include <linux/pci_ids.h>
#include <linux/slab.h>
#include <linux/delay.h>
#include <linux/dmi.h>
#include <linux/edac.h>
#include <linux/mmzone.h>
#include <linux/smp.h>
#include <asm/mce.h>
#include <asm/processor.h>
#include <asm/div64.h>
#include "edac_core.h"
/* Static vars */
static LIST_HEAD(i7core_edac_list);
static DEFINE_MUTEX(i7core_edac_lock);
static int probed;
static int use_pci_fixup;
module_param(use_pci_fixup, int, 0444);
MODULE_PARM_DESC(use_pci_fixup, "Enable PCI fixup to seek for hidden devices");
/*
* This is used for Nehalem-EP and Nehalem-EX devices, where the non-core
* registers start at bus 255, and are not reported by BIOS.
* We currently find devices with only 2 sockets. In order to support more QPI
* Quick Path Interconnect, just increment this number.
*/
#define MAX_SOCKET_BUSES 2
/*
* Alter this version for the module when modifications are made
*/
#define I7CORE_REVISION " Ver: 1.0.0"
#define EDAC_MOD_STR "i7core_edac"
/*
* Debug macros
*/
#define i7core_printk(level, fmt, arg...) \
edac_printk(level, "i7core", fmt, ##arg)
#define i7core_mc_printk(mci, level, fmt, arg...) \
edac_mc_chipset_printk(mci, level, "i7core", fmt, ##arg)
/*
* i7core Memory Controller Registers
*/
/* OFFSETS for Device 0 Function 0 */
#define MC_CFG_CONTROL 0x90
#define MC_CFG_UNLOCK 0x02
#define MC_CFG_LOCK 0x00
/* OFFSETS for Device 3 Function 0 */
#define MC_CONTROL 0x48
#define MC_STATUS 0x4c
#define MC_MAX_DOD 0x64
/*
* OFFSETS for Device 3 Function 4, as indicated on Xeon 5500 datasheet:
* http://www.arrownac.com/manufacturers/intel/s/nehalem/5500-datasheet-v2.pdf
*/
#define MC_TEST_ERR_RCV1 0x60
#define DIMM2_COR_ERR(r) ((r) & 0x7fff)
#define MC_TEST_ERR_RCV0 0x64
#define DIMM1_COR_ERR(r) (((r) >> 16) & 0x7fff)
#define DIMM0_COR_ERR(r) ((r) & 0x7fff)
/* OFFSETS for Device 3 Function 2, as indicated on Xeon 5500 datasheet */
#define MC_SSRCONTROL 0x48
#define SSR_MODE_DISABLE 0x00
#define SSR_MODE_ENABLE 0x01
#define SSR_MODE_MASK 0x03
#define MC_SCRUB_CONTROL 0x4c
#define STARTSCRUB (1 << 24)
#define SCRUBINTERVAL_MASK 0xffffff
#define MC_COR_ECC_CNT_0 0x80
#define MC_COR_ECC_CNT_1 0x84
#define MC_COR_ECC_CNT_2 0x88
#define MC_COR_ECC_CNT_3 0x8c
#define MC_COR_ECC_CNT_4 0x90
#define MC_COR_ECC_CNT_5 0x94
#define DIMM_TOP_COR_ERR(r) (((r) >> 16) & 0x7fff)
#define DIMM_BOT_COR_ERR(r) ((r) & 0x7fff)
/* OFFSETS for Devices 4,5 and 6 Function 0 */
#define MC_CHANNEL_DIMM_INIT_PARAMS 0x58
#define THREE_DIMMS_PRESENT (1 << 24)
#define SINGLE_QUAD_RANK_PRESENT (1 << 23)
#define QUAD_RANK_PRESENT (1 << 22)
#define REGISTERED_DIMM (1 << 15)
#define MC_CHANNEL_MAPPER 0x60
#define RDLCH(r, ch) ((((r) >> (3 + (ch * 6))) & 0x07) - 1)
#define WRLCH(r, ch) ((((r) >> (ch * 6)) & 0x07) - 1)
#define MC_CHANNEL_RANK_PRESENT 0x7c
#define RANK_PRESENT_MASK 0xffff
#define MC_CHANNEL_ADDR_MATCH 0xf0
#define MC_CHANNEL_ERROR_MASK 0xf8
#define MC_CHANNEL_ERROR_INJECT 0xfc
#define INJECT_ADDR_PARITY 0x10
#define INJECT_ECC 0x08
#define MASK_CACHELINE 0x06
#define MASK_FULL_CACHELINE 0x06
#define MASK_MSB32_CACHELINE 0x04
#define MASK_LSB32_CACHELINE 0x02
#define NO_MASK_CACHELINE 0x00
#define REPEAT_EN 0x01
/* OFFSETS for Devices 4,5 and 6 Function 1 */
#define MC_DOD_CH_DIMM0 0x48
#define MC_DOD_CH_DIMM1 0x4c
#define MC_DOD_CH_DIMM2 0x50
#define RANKOFFSET_MASK ((1 << 12) | (1 << 11) | (1 << 10))
#define RANKOFFSET(x) ((x & RANKOFFSET_MASK) >> 10)
#define DIMM_PRESENT_MASK (1 << 9)
#define DIMM_PRESENT(x) (((x) & DIMM_PRESENT_MASK) >> 9)
#define MC_DOD_NUMBANK_MASK ((1 << 8) | (1 << 7))
#define MC_DOD_NUMBANK(x) (((x) & MC_DOD_NUMBANK_MASK) >> 7)
#define MC_DOD_NUMRANK_MASK ((1 << 6) | (1 << 5))
#define MC_DOD_NUMRANK(x) (((x) & MC_DOD_NUMRANK_MASK) >> 5)
#define MC_DOD_NUMROW_MASK ((1 << 4) | (1 << 3) | (1 << 2))
#define MC_DOD_NUMROW(x) (((x) & MC_DOD_NUMROW_MASK) >> 2)
#define MC_DOD_NUMCOL_MASK 3
#define MC_DOD_NUMCOL(x) ((x) & MC_DOD_NUMCOL_MASK)
#define MC_RANK_PRESENT 0x7c
#define MC_SAG_CH_0 0x80
#define MC_SAG_CH_1 0x84
#define MC_SAG_CH_2 0x88
#define MC_SAG_CH_3 0x8c
#define MC_SAG_CH_4 0x90
#define MC_SAG_CH_5 0x94
#define MC_SAG_CH_6 0x98
#define MC_SAG_CH_7 0x9c
#define MC_RIR_LIMIT_CH_0 0x40
#define MC_RIR_LIMIT_CH_1 0x44
#define MC_RIR_LIMIT_CH_2 0x48
#define MC_RIR_LIMIT_CH_3 0x4C
#define MC_RIR_LIMIT_CH_4 0x50
#define MC_RIR_LIMIT_CH_5 0x54
#define MC_RIR_LIMIT_CH_6 0x58
#define MC_RIR_LIMIT_CH_7 0x5C
#define MC_RIR_LIMIT_MASK ((1 << 10) - 1)
#define MC_RIR_WAY_CH 0x80
#define MC_RIR_WAY_OFFSET_MASK (((1 << 14) - 1) & ~0x7)
#define MC_RIR_WAY_RANK_MASK 0x7
/*
* i7core structs
*/
#define NUM_CHANS 3
#define MAX_DIMMS 3 /* Max DIMMS per channel */
#define MAX_MCR_FUNC 4
#define MAX_CHAN_FUNC 3
struct i7core_info {
u32 mc_control;
u32 mc_status;
u32 max_dod;
u32 ch_map;
};
struct i7core_inject {
int enable;
u32 section;
u32 type;
u32 eccmask;
/* Error address mask */
int channel, dimm, rank, bank, page, col;
};
struct i7core_channel {
bool is_3dimms_present;
bool is_single_4rank;
bool has_4rank;
u32 dimms;
};
struct pci_id_descr {
int dev;
int func;
int dev_id;
int optional;
};
struct pci_id_table {
const struct pci_id_descr *descr;
int n_devs;
};
struct i7core_dev {
struct list_head list;
u8 socket;
struct pci_dev **pdev;
int n_devs;
struct mem_ctl_info *mci;
};
struct i7core_pvt {
struct device *addrmatch_dev, *chancounts_dev;
struct pci_dev *pci_noncore;
struct pci_dev *pci_mcr[MAX_MCR_FUNC + 1];
struct pci_dev *pci_ch[NUM_CHANS][MAX_CHAN_FUNC + 1];
struct i7core_dev *i7core_dev;
struct i7core_info info;
struct i7core_inject inject;
struct i7core_channel channel[NUM_CHANS];
int ce_count_available;
/* ECC corrected errors counts per udimm */
unsigned long udimm_ce_count[MAX_DIMMS];
int udimm_last_ce_count[MAX_DIMMS];
/* ECC corrected errors counts per rdimm */
unsigned long rdimm_ce_count[NUM_CHANS][MAX_DIMMS];
int rdimm_last_ce_count[NUM_CHANS][MAX_DIMMS];
bool is_registered, enable_scrub;
/* Fifo double buffers */
struct mce mce_entry[MCE_LOG_LEN];
struct mce mce_outentry[MCE_LOG_LEN];
/* Fifo in/out counters */
unsigned mce_in, mce_out;
/* Count indicator to show errors not got */
unsigned mce_overrun;
/* DCLK Frequency used for computing scrub rate */
int dclk_freq;
/* Struct to control EDAC polling */
struct edac_pci_ctl_info *i7core_pci;
};
#define PCI_DESCR(device, function, device_id) \
.dev = (device), \
.func = (function), \
.dev_id = (device_id)
static const struct pci_id_descr pci_dev_descr_i7core_nehalem[] = {
/* Memory controller */
{ PCI_DESCR(3, 0, PCI_DEVICE_ID_INTEL_I7_MCR) },
{ PCI_DESCR(3, 1, PCI_DEVICE_ID_INTEL_I7_MC_TAD) },
/* Exists only for RDIMM */
{ PCI_DESCR(3, 2, PCI_DEVICE_ID_INTEL_I7_MC_RAS), .optional = 1 },
{ PCI_DESCR(3, 4, PCI_DEVICE_ID_INTEL_I7_MC_TEST) },
/* Channel 0 */
{ PCI_DESCR(4, 0, PCI_DEVICE_ID_INTEL_I7_MC_CH0_CTRL) },
{ PCI_DESCR(4, 1, PCI_DEVICE_ID_INTEL_I7_MC_CH0_ADDR) },
{ PCI_DESCR(4, 2, PCI_DEVICE_ID_INTEL_I7_MC_CH0_RANK) },
{ PCI_DESCR(4, 3, PCI_DEVICE_ID_INTEL_I7_MC_CH0_TC) },
/* Channel 1 */
{ PCI_DESCR(5, 0, PCI_DEVICE_ID_INTEL_I7_MC_CH1_CTRL) },
{ PCI_DESCR(5, 1, PCI_DEVICE_ID_INTEL_I7_MC_CH1_ADDR) },
{ PCI_DESCR(5, 2, PCI_DEVICE_ID_INTEL_I7_MC_CH1_RANK) },
{ PCI_DESCR(5, 3, PCI_DEVICE_ID_INTEL_I7_MC_CH1_TC) },
/* Channel 2 */
{ PCI_DESCR(6, 0, PCI_DEVICE_ID_INTEL_I7_MC_CH2_CTRL) },
{ PCI_DESCR(6, 1, PCI_DEVICE_ID_INTEL_I7_MC_CH2_ADDR) },
{ PCI_DESCR(6, 2, PCI_DEVICE_ID_INTEL_I7_MC_CH2_RANK) },
{ PCI_DESCR(6, 3, PCI_DEVICE_ID_INTEL_I7_MC_CH2_TC) },
/* Generic Non-core registers */
/*
* This is the PCI device on i7core and on Xeon 35xx (8086:2c41)
* On Xeon 55xx, however, it has a different id (8086:2c40). So,
* the probing code needs to test for the other address in case of
* failure of this one
*/
{ PCI_DESCR(0, 0, PCI_DEVICE_ID_INTEL_I7_NONCORE) },
};
static const struct pci_id_descr pci_dev_descr_lynnfield[] = {
{ PCI_DESCR( 3, 0, PCI_DEVICE_ID_INTEL_LYNNFIELD_MCR) },
{ PCI_DESCR( 3, 1, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_TAD) },
{ PCI_DESCR( 3, 4, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_TEST) },
{ PCI_DESCR( 4, 0, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH0_CTRL) },
{ PCI_DESCR( 4, 1, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH0_ADDR) },
{ PCI_DESCR( 4, 2, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH0_RANK) },
{ PCI_DESCR( 4, 3, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH0_TC) },
{ PCI_DESCR( 5, 0, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH1_CTRL) },
{ PCI_DESCR( 5, 1, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH1_ADDR) },
{ PCI_DESCR( 5, 2, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH1_RANK) },
{ PCI_DESCR( 5, 3, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH1_TC) },
/*
* This is the PCI device has an alternate address on some
* processors like Core i7 860
*/
{ PCI_DESCR( 0, 0, PCI_DEVICE_ID_INTEL_LYNNFIELD_NONCORE) },
};
static const struct pci_id_descr pci_dev_descr_i7core_westmere[] = {
/* Memory controller */
{ PCI_DESCR(3, 0, PCI_DEVICE_ID_INTEL_LYNNFIELD_MCR_REV2) },
{ PCI_DESCR(3, 1, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_TAD_REV2) },
/* Exists only for RDIMM */
{ PCI_DESCR(3, 2, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_RAS_REV2), .optional = 1 },
{ PCI_DESCR(3, 4, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_TEST_REV2) },
/* Channel 0 */
{ PCI_DESCR(4, 0, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH0_CTRL_REV2) },
{ PCI_DESCR(4, 1, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH0_ADDR_REV2) },
{ PCI_DESCR(4, 2, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH0_RANK_REV2) },
{ PCI_DESCR(4, 3, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH0_TC_REV2) },
/* Channel 1 */
{ PCI_DESCR(5, 0, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH1_CTRL_REV2) },
{ PCI_DESCR(5, 1, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH1_ADDR_REV2) },
{ PCI_DESCR(5, 2, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH1_RANK_REV2) },
{ PCI_DESCR(5, 3, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH1_TC_REV2) },
/* Channel 2 */
{ PCI_DESCR(6, 0, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH2_CTRL_REV2) },
{ PCI_DESCR(6, 1, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH2_ADDR_REV2) },
{ PCI_DESCR(6, 2, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH2_RANK_REV2) },
{ PCI_DESCR(6, 3, PCI_DEVICE_ID_INTEL_LYNNFIELD_MC_CH2_TC_REV2) },
/* Generic Non-core registers */
{ PCI_DESCR(0, 0, PCI_DEVICE_ID_INTEL_LYNNFIELD_NONCORE_REV2) },
};
#define PCI_ID_TABLE_ENTRY(A) { .descr=A, .n_devs = ARRAY_SIZE(A) }
static const struct pci_id_table pci_dev_table[] = {
PCI_ID_TABLE_ENTRY(pci_dev_descr_i7core_nehalem),
PCI_ID_TABLE_ENTRY(pci_dev_descr_lynnfield),
PCI_ID_TABLE_ENTRY(pci_dev_descr_i7core_westmere),
{0,} /* 0 terminated list. */
};
/*
* pci_device_id table for which devices we are looking for
*/
static const struct pci_device_id i7core_pci_tbl[] = {
{PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_X58_HUB_MGMT)},
{PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_LYNNFIELD_QPI_LINK0)},
{0,} /* 0 terminated list. */
};
/****************************************************************************
Ancillary status routines
****************************************************************************/
/* MC_CONTROL bits */
#define CH_ACTIVE(pvt, ch) ((pvt)->info.mc_control & (1 << (8 + ch)))
#define ECCx8(pvt) ((pvt)->info.mc_control & (1 << 1))
/* MC_STATUS bits */
#define ECC_ENABLED(pvt) ((pvt)->info.mc_status & (1 << 4))
#define CH_DISABLED(pvt, ch) ((pvt)->info.mc_status & (1 << ch))
/* MC_MAX_DOD read functions */
static inline int numdimms(u32 dimms)
{
return (dimms & 0x3) + 1;
}
static inline int numrank(u32 rank)
{
static const int ranks[] = { 1, 2, 4, -EINVAL };
return ranks[rank & 0x3];
}
static inline int numbank(u32 bank)
{
static const int banks[] = { 4, 8, 16, -EINVAL };
return banks[bank & 0x3];
}
static inline int numrow(u32 row)
{
static const int rows[] = {
1 << 12, 1 << 13, 1 << 14, 1 << 15,
1 << 16, -EINVAL, -EINVAL, -EINVAL,
};
return rows[row & 0x7];
}
static inline int numcol(u32 col)
{
static const int cols[] = {
1 << 10, 1 << 11, 1 << 12, -EINVAL,
};
return cols[col & 0x3];
}
static struct i7core_dev *get_i7core_dev(u8 socket)
{
struct i7core_dev *i7core_dev;
list_for_each_entry(i7core_dev, &i7core_edac_list, list) {
if (i7core_dev->socket == socket)
return i7core_dev;
}
return NULL;
}
static struct i7core_dev *alloc_i7core_dev(u8 socket,
const struct pci_id_table *table)
{
struct i7core_dev *i7core_dev;
i7core_dev = kzalloc(sizeof(*i7core_dev), GFP_KERNEL);
if (!i7core_dev)
return NULL;
i7core_dev->pdev = kzalloc(sizeof(*i7core_dev->pdev) * table->n_devs,
GFP_KERNEL);
if (!i7core_dev->pdev) {
kfree(i7core_dev);
return NULL;
}
i7core_dev->socket = socket;
i7core_dev->n_devs = table->n_devs;
list_add_tail(&i7core_dev->list, &i7core_edac_list);
return i7core_dev;
}
static void free_i7core_dev(struct i7core_dev *i7core_dev)
{
list_del(&i7core_dev->list);
kfree(i7core_dev->pdev);
kfree(i7core_dev);
}
/****************************************************************************
Memory check routines
****************************************************************************/
edac: move dimm properties to struct dimm_info On systems based on chip select rows, all channels need to use memories with the same properties, otherwise the memories on channels A and B won't be recognized. However, such assumption is not true for all types of memory controllers. Controllers for FB-DIMM's don't have such requirements. Also, modern Intel controllers seem to be capable of handling such differences. So, we need to get rid of storing the DIMM information into a per-csrow data, storing it, instead at the right place. The first step is to move grain, mtype, dtype and edac_mode to the per-dimm struct. Reviewed-by: Aristeu Rozanski <arozansk@redhat.com> Reviewed-by: Borislav Petkov <borislav.petkov@amd.com> Acked-by: Chris Metcalf <cmetcalf@tilera.com> Cc: Doug Thompson <norsk5@yahoo.com> Cc: Borislav Petkov <borislav.petkov@amd.com> Cc: Mark Gross <mark.gross@intel.com> Cc: Jason Uhlenkott <juhlenko@akamai.com> Cc: Tim Small <tim@buttersideup.com> Cc: Ranganathan Desikan <ravi@jetztechnologies.com> Cc: "Arvind R." <arvino55@gmail.com> Cc: Olof Johansson <olof@lixom.net> Cc: Egor Martovetsky <egor@pasemi.com> Cc: Michal Marek <mmarek@suse.cz> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Joe Perches <joe@perches.com> Cc: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Hitoshi Mitake <h.mitake@gmail.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: James Bottomley <James.Bottomley@parallels.com> Cc: "Niklas Söderlund" <niklas.soderlund@ericsson.com> Cc: Shaohui Xie <Shaohui.Xie@freescale.com> Cc: Josh Boyer <jwboyer@gmail.com> Cc: Mike Williams <mike@mikebwilliams.com> Cc: linuxppc-dev@lists.ozlabs.org Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2012-01-28 04:38:08 +07:00
static int get_dimm_config(struct mem_ctl_info *mci)
{
struct i7core_pvt *pvt = mci->pvt_info;
struct pci_dev *pdev;
int i, j;
enum edac_type mode;
enum mem_type mtype;
edac: move dimm properties to struct dimm_info On systems based on chip select rows, all channels need to use memories with the same properties, otherwise the memories on channels A and B won't be recognized. However, such assumption is not true for all types of memory controllers. Controllers for FB-DIMM's don't have such requirements. Also, modern Intel controllers seem to be capable of handling such differences. So, we need to get rid of storing the DIMM information into a per-csrow data, storing it, instead at the right place. The first step is to move grain, mtype, dtype and edac_mode to the per-dimm struct. Reviewed-by: Aristeu Rozanski <arozansk@redhat.com> Reviewed-by: Borislav Petkov <borislav.petkov@amd.com> Acked-by: Chris Metcalf <cmetcalf@tilera.com> Cc: Doug Thompson <norsk5@yahoo.com> Cc: Borislav Petkov <borislav.petkov@amd.com> Cc: Mark Gross <mark.gross@intel.com> Cc: Jason Uhlenkott <juhlenko@akamai.com> Cc: Tim Small <tim@buttersideup.com> Cc: Ranganathan Desikan <ravi@jetztechnologies.com> Cc: "Arvind R." <arvino55@gmail.com> Cc: Olof Johansson <olof@lixom.net> Cc: Egor Martovetsky <egor@pasemi.com> Cc: Michal Marek <mmarek@suse.cz> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Joe Perches <joe@perches.com> Cc: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Hitoshi Mitake <h.mitake@gmail.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: James Bottomley <James.Bottomley@parallels.com> Cc: "Niklas Söderlund" <niklas.soderlund@ericsson.com> Cc: Shaohui Xie <Shaohui.Xie@freescale.com> Cc: Josh Boyer <jwboyer@gmail.com> Cc: Mike Williams <mike@mikebwilliams.com> Cc: linuxppc-dev@lists.ozlabs.org Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2012-01-28 04:38:08 +07:00
struct dimm_info *dimm;
/* Get data from the MC register, function 0 */
pdev = pvt->pci_mcr[0];
if (!pdev)
return -ENODEV;
/* Device 3 function 0 reads */
pci_read_config_dword(pdev, MC_CONTROL, &pvt->info.mc_control);
pci_read_config_dword(pdev, MC_STATUS, &pvt->info.mc_status);
pci_read_config_dword(pdev, MC_MAX_DOD, &pvt->info.max_dod);
pci_read_config_dword(pdev, MC_CHANNEL_MAPPER, &pvt->info.ch_map);
edac_dbg(0, "QPI %d control=0x%08x status=0x%08x dod=0x%08x map=0x%08x\n",
pvt->i7core_dev->socket, pvt->info.mc_control,
pvt->info.mc_status, pvt->info.max_dod, pvt->info.ch_map);
if (ECC_ENABLED(pvt)) {
edac_dbg(0, "ECC enabled with x%d SDCC\n", ECCx8(pvt) ? 8 : 4);
if (ECCx8(pvt))
mode = EDAC_S8ECD8ED;
else
mode = EDAC_S4ECD4ED;
} else {
edac_dbg(0, "ECC disabled\n");
mode = EDAC_NONE;
}
/* FIXME: need to handle the error codes */
edac_dbg(0, "DOD Max limits: DIMMS: %d, %d-ranked, %d-banked x%x x 0x%x\n",
numdimms(pvt->info.max_dod),
numrank(pvt->info.max_dod >> 2),
numbank(pvt->info.max_dod >> 4),
numrow(pvt->info.max_dod >> 6),
numcol(pvt->info.max_dod >> 9));
for (i = 0; i < NUM_CHANS; i++) {
u32 data, dimm_dod[3], value[8];
if (!pvt->pci_ch[i][0])
continue;
if (!CH_ACTIVE(pvt, i)) {
edac_dbg(0, "Channel %i is not active\n", i);
continue;
}
if (CH_DISABLED(pvt, i)) {
edac_dbg(0, "Channel %i is disabled\n", i);
continue;
}
/* Devices 4-6 function 0 */
pci_read_config_dword(pvt->pci_ch[i][0],
MC_CHANNEL_DIMM_INIT_PARAMS, &data);
if (data & THREE_DIMMS_PRESENT)
pvt->channel[i].is_3dimms_present = true;
if (data & SINGLE_QUAD_RANK_PRESENT)
pvt->channel[i].is_single_4rank = true;
if (data & QUAD_RANK_PRESENT)
pvt->channel[i].has_4rank = true;
if (data & REGISTERED_DIMM)
mtype = MEM_RDDR3;
else
mtype = MEM_DDR3;
/* Devices 4-6 function 1 */
pci_read_config_dword(pvt->pci_ch[i][1],
MC_DOD_CH_DIMM0, &dimm_dod[0]);
pci_read_config_dword(pvt->pci_ch[i][1],
MC_DOD_CH_DIMM1, &dimm_dod[1]);
pci_read_config_dword(pvt->pci_ch[i][1],
MC_DOD_CH_DIMM2, &dimm_dod[2]);
edac_dbg(0, "Ch%d phy rd%d, wr%d (0x%08x): %s%s%s%cDIMMs\n",
i,
RDLCH(pvt->info.ch_map, i), WRLCH(pvt->info.ch_map, i),
data,
pvt->channel[i].is_3dimms_present ? "3DIMMS " : "",
pvt->channel[i].is_3dimms_present ? "SINGLE_4R " : "",
pvt->channel[i].has_4rank ? "HAS_4R " : "",
(data & REGISTERED_DIMM) ? 'R' : 'U');
for (j = 0; j < 3; j++) {
u32 banks, ranks, rows, cols;
u32 size, npages;
if (!DIMM_PRESENT(dimm_dod[j]))
continue;
dimm = EDAC_DIMM_PTR(mci->layers, mci->dimms, mci->n_layers,
i, j, 0);
banks = numbank(MC_DOD_NUMBANK(dimm_dod[j]));
ranks = numrank(MC_DOD_NUMRANK(dimm_dod[j]));
rows = numrow(MC_DOD_NUMROW(dimm_dod[j]));
cols = numcol(MC_DOD_NUMCOL(dimm_dod[j]));
/* DDR3 has 8 I/O banks */
size = (rows * cols * banks * ranks) >> (20 - 3);
edac_dbg(0, "\tdimm %d %d Mb offset: %x, bank: %d, rank: %d, row: %#x, col: %#x\n",
j, size,
RANKOFFSET(dimm_dod[j]),
banks, ranks, rows, cols);
npages = MiB_TO_PAGES(size);
dimm->nr_pages = npages;
switch (banks) {
case 4:
edac: move dimm properties to struct dimm_info On systems based on chip select rows, all channels need to use memories with the same properties, otherwise the memories on channels A and B won't be recognized. However, such assumption is not true for all types of memory controllers. Controllers for FB-DIMM's don't have such requirements. Also, modern Intel controllers seem to be capable of handling such differences. So, we need to get rid of storing the DIMM information into a per-csrow data, storing it, instead at the right place. The first step is to move grain, mtype, dtype and edac_mode to the per-dimm struct. Reviewed-by: Aristeu Rozanski <arozansk@redhat.com> Reviewed-by: Borislav Petkov <borislav.petkov@amd.com> Acked-by: Chris Metcalf <cmetcalf@tilera.com> Cc: Doug Thompson <norsk5@yahoo.com> Cc: Borislav Petkov <borislav.petkov@amd.com> Cc: Mark Gross <mark.gross@intel.com> Cc: Jason Uhlenkott <juhlenko@akamai.com> Cc: Tim Small <tim@buttersideup.com> Cc: Ranganathan Desikan <ravi@jetztechnologies.com> Cc: "Arvind R." <arvino55@gmail.com> Cc: Olof Johansson <olof@lixom.net> Cc: Egor Martovetsky <egor@pasemi.com> Cc: Michal Marek <mmarek@suse.cz> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Joe Perches <joe@perches.com> Cc: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Hitoshi Mitake <h.mitake@gmail.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: James Bottomley <James.Bottomley@parallels.com> Cc: "Niklas Söderlund" <niklas.soderlund@ericsson.com> Cc: Shaohui Xie <Shaohui.Xie@freescale.com> Cc: Josh Boyer <jwboyer@gmail.com> Cc: Mike Williams <mike@mikebwilliams.com> Cc: linuxppc-dev@lists.ozlabs.org Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2012-01-28 04:38:08 +07:00
dimm->dtype = DEV_X4;
break;
case 8:
edac: move dimm properties to struct dimm_info On systems based on chip select rows, all channels need to use memories with the same properties, otherwise the memories on channels A and B won't be recognized. However, such assumption is not true for all types of memory controllers. Controllers for FB-DIMM's don't have such requirements. Also, modern Intel controllers seem to be capable of handling such differences. So, we need to get rid of storing the DIMM information into a per-csrow data, storing it, instead at the right place. The first step is to move grain, mtype, dtype and edac_mode to the per-dimm struct. Reviewed-by: Aristeu Rozanski <arozansk@redhat.com> Reviewed-by: Borislav Petkov <borislav.petkov@amd.com> Acked-by: Chris Metcalf <cmetcalf@tilera.com> Cc: Doug Thompson <norsk5@yahoo.com> Cc: Borislav Petkov <borislav.petkov@amd.com> Cc: Mark Gross <mark.gross@intel.com> Cc: Jason Uhlenkott <juhlenko@akamai.com> Cc: Tim Small <tim@buttersideup.com> Cc: Ranganathan Desikan <ravi@jetztechnologies.com> Cc: "Arvind R." <arvino55@gmail.com> Cc: Olof Johansson <olof@lixom.net> Cc: Egor Martovetsky <egor@pasemi.com> Cc: Michal Marek <mmarek@suse.cz> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Joe Perches <joe@perches.com> Cc: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Hitoshi Mitake <h.mitake@gmail.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: James Bottomley <James.Bottomley@parallels.com> Cc: "Niklas Söderlund" <niklas.soderlund@ericsson.com> Cc: Shaohui Xie <Shaohui.Xie@freescale.com> Cc: Josh Boyer <jwboyer@gmail.com> Cc: Mike Williams <mike@mikebwilliams.com> Cc: linuxppc-dev@lists.ozlabs.org Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2012-01-28 04:38:08 +07:00
dimm->dtype = DEV_X8;
break;
case 16:
edac: move dimm properties to struct dimm_info On systems based on chip select rows, all channels need to use memories with the same properties, otherwise the memories on channels A and B won't be recognized. However, such assumption is not true for all types of memory controllers. Controllers for FB-DIMM's don't have such requirements. Also, modern Intel controllers seem to be capable of handling such differences. So, we need to get rid of storing the DIMM information into a per-csrow data, storing it, instead at the right place. The first step is to move grain, mtype, dtype and edac_mode to the per-dimm struct. Reviewed-by: Aristeu Rozanski <arozansk@redhat.com> Reviewed-by: Borislav Petkov <borislav.petkov@amd.com> Acked-by: Chris Metcalf <cmetcalf@tilera.com> Cc: Doug Thompson <norsk5@yahoo.com> Cc: Borislav Petkov <borislav.petkov@amd.com> Cc: Mark Gross <mark.gross@intel.com> Cc: Jason Uhlenkott <juhlenko@akamai.com> Cc: Tim Small <tim@buttersideup.com> Cc: Ranganathan Desikan <ravi@jetztechnologies.com> Cc: "Arvind R." <arvino55@gmail.com> Cc: Olof Johansson <olof@lixom.net> Cc: Egor Martovetsky <egor@pasemi.com> Cc: Michal Marek <mmarek@suse.cz> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Joe Perches <joe@perches.com> Cc: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Hitoshi Mitake <h.mitake@gmail.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: James Bottomley <James.Bottomley@parallels.com> Cc: "Niklas Söderlund" <niklas.soderlund@ericsson.com> Cc: Shaohui Xie <Shaohui.Xie@freescale.com> Cc: Josh Boyer <jwboyer@gmail.com> Cc: Mike Williams <mike@mikebwilliams.com> Cc: linuxppc-dev@lists.ozlabs.org Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2012-01-28 04:38:08 +07:00
dimm->dtype = DEV_X16;
break;
default:
edac: move dimm properties to struct dimm_info On systems based on chip select rows, all channels need to use memories with the same properties, otherwise the memories on channels A and B won't be recognized. However, such assumption is not true for all types of memory controllers. Controllers for FB-DIMM's don't have such requirements. Also, modern Intel controllers seem to be capable of handling such differences. So, we need to get rid of storing the DIMM information into a per-csrow data, storing it, instead at the right place. The first step is to move grain, mtype, dtype and edac_mode to the per-dimm struct. Reviewed-by: Aristeu Rozanski <arozansk@redhat.com> Reviewed-by: Borislav Petkov <borislav.petkov@amd.com> Acked-by: Chris Metcalf <cmetcalf@tilera.com> Cc: Doug Thompson <norsk5@yahoo.com> Cc: Borislav Petkov <borislav.petkov@amd.com> Cc: Mark Gross <mark.gross@intel.com> Cc: Jason Uhlenkott <juhlenko@akamai.com> Cc: Tim Small <tim@buttersideup.com> Cc: Ranganathan Desikan <ravi@jetztechnologies.com> Cc: "Arvind R." <arvino55@gmail.com> Cc: Olof Johansson <olof@lixom.net> Cc: Egor Martovetsky <egor@pasemi.com> Cc: Michal Marek <mmarek@suse.cz> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Joe Perches <joe@perches.com> Cc: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Hitoshi Mitake <h.mitake@gmail.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: James Bottomley <James.Bottomley@parallels.com> Cc: "Niklas Söderlund" <niklas.soderlund@ericsson.com> Cc: Shaohui Xie <Shaohui.Xie@freescale.com> Cc: Josh Boyer <jwboyer@gmail.com> Cc: Mike Williams <mike@mikebwilliams.com> Cc: linuxppc-dev@lists.ozlabs.org Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2012-01-28 04:38:08 +07:00
dimm->dtype = DEV_UNKNOWN;
}
edac: move dimm properties to struct dimm_info On systems based on chip select rows, all channels need to use memories with the same properties, otherwise the memories on channels A and B won't be recognized. However, such assumption is not true for all types of memory controllers. Controllers for FB-DIMM's don't have such requirements. Also, modern Intel controllers seem to be capable of handling such differences. So, we need to get rid of storing the DIMM information into a per-csrow data, storing it, instead at the right place. The first step is to move grain, mtype, dtype and edac_mode to the per-dimm struct. Reviewed-by: Aristeu Rozanski <arozansk@redhat.com> Reviewed-by: Borislav Petkov <borislav.petkov@amd.com> Acked-by: Chris Metcalf <cmetcalf@tilera.com> Cc: Doug Thompson <norsk5@yahoo.com> Cc: Borislav Petkov <borislav.petkov@amd.com> Cc: Mark Gross <mark.gross@intel.com> Cc: Jason Uhlenkott <juhlenko@akamai.com> Cc: Tim Small <tim@buttersideup.com> Cc: Ranganathan Desikan <ravi@jetztechnologies.com> Cc: "Arvind R." <arvino55@gmail.com> Cc: Olof Johansson <olof@lixom.net> Cc: Egor Martovetsky <egor@pasemi.com> Cc: Michal Marek <mmarek@suse.cz> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Joe Perches <joe@perches.com> Cc: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Hitoshi Mitake <h.mitake@gmail.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: James Bottomley <James.Bottomley@parallels.com> Cc: "Niklas Söderlund" <niklas.soderlund@ericsson.com> Cc: Shaohui Xie <Shaohui.Xie@freescale.com> Cc: Josh Boyer <jwboyer@gmail.com> Cc: Mike Williams <mike@mikebwilliams.com> Cc: linuxppc-dev@lists.ozlabs.org Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2012-01-28 04:38:08 +07:00
snprintf(dimm->label, sizeof(dimm->label),
"CPU#%uChannel#%u_DIMM#%u",
pvt->i7core_dev->socket, i, j);
dimm->grain = 8;
dimm->edac_mode = mode;
dimm->mtype = mtype;
}
pci_read_config_dword(pdev, MC_SAG_CH_0, &value[0]);
pci_read_config_dword(pdev, MC_SAG_CH_1, &value[1]);
pci_read_config_dword(pdev, MC_SAG_CH_2, &value[2]);
pci_read_config_dword(pdev, MC_SAG_CH_3, &value[3]);
pci_read_config_dword(pdev, MC_SAG_CH_4, &value[4]);
pci_read_config_dword(pdev, MC_SAG_CH_5, &value[5]);
pci_read_config_dword(pdev, MC_SAG_CH_6, &value[6]);
pci_read_config_dword(pdev, MC_SAG_CH_7, &value[7]);
edac_dbg(1, "\t[%i] DIVBY3\tREMOVED\tOFFSET\n", i);
for (j = 0; j < 8; j++)
edac_dbg(1, "\t\t%#x\t%#x\t%#x\n",
(value[j] >> 27) & 0x1,
(value[j] >> 24) & 0x7,
(value[j] & ((1 << 24) - 1)));
}
return 0;
}
/****************************************************************************
Error insertion routines
****************************************************************************/
#define to_mci(k) container_of(k, struct mem_ctl_info, dev)
/* The i7core has independent error injection features per channel.
However, to have a simpler code, we don't allow enabling error injection
on more than one channel.
Also, since a change at an inject parameter will be applied only at enable,
we're disabling error injection on all write calls to the sysfs nodes that
controls the error code injection.
*/
static int disable_inject(const struct mem_ctl_info *mci)
{
struct i7core_pvt *pvt = mci->pvt_info;
pvt->inject.enable = 0;
if (!pvt->pci_ch[pvt->inject.channel][0])
return -ENODEV;
pci_write_config_dword(pvt->pci_ch[pvt->inject.channel][0],
MC_CHANNEL_ERROR_INJECT, 0);
return 0;
}
/*
* i7core inject inject.section
*
* accept and store error injection inject.section value
* bit 0 - refers to the lower 32-byte half cacheline
* bit 1 - refers to the upper 32-byte half cacheline
*/
static ssize_t i7core_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 i7core_pvt *pvt = mci->pvt_info;
unsigned long value;
int rc;
if (pvt->inject.enable)
disable_inject(mci);
rc = kstrtoul(data, 10, &value);
if ((rc < 0) || (value > 3))
return -EIO;
pvt->inject.section = (u32) value;
return count;
}
static ssize_t i7core_inject_section_show(struct device *dev,
struct device_attribute *mattr,
char *data)
{
struct mem_ctl_info *mci = to_mci(dev);
struct i7core_pvt *pvt = mci->pvt_info;
return sprintf(data, "0x%08x\n", pvt->inject.section);
}
/*
* i7core inject.type
*
* accept and store error injection inject.section value
* bit 0 - repeat enable - Enable error repetition
* bit 1 - inject ECC error
* bit 2 - inject parity error
*/
static ssize_t i7core_inject_type_store(struct device *dev,
struct device_attribute *mattr,
const char *data, size_t count)
{
struct mem_ctl_info *mci = to_mci(dev);
struct i7core_pvt *pvt = mci->pvt_info;
unsigned long value;
int rc;
if (pvt->inject.enable)
disable_inject(mci);
rc = kstrtoul(data, 10, &value);
if ((rc < 0) || (value > 7))
return -EIO;
pvt->inject.type = (u32) value;
return count;
}
static ssize_t i7core_inject_type_show(struct device *dev,
struct device_attribute *mattr,
char *data)
{
struct mem_ctl_info *mci = to_mci(dev);
struct i7core_pvt *pvt = mci->pvt_info;
return sprintf(data, "0x%08x\n", pvt->inject.type);
}
/*
* i7core_inject_inject.eccmask_store
*
* The type of error (UE/CE) will depend on the inject.eccmask value:
* Any bits set to a 1 will flip the corresponding ECC bit
* Correctable errors can be injected by flipping 1 bit or the bits within
* a symbol pair (2 consecutive aligned 8-bit pairs - i.e. 7:0 and 15:8 or
* 23:16 and 31:24). Flipping bits in two symbol pairs will cause an
* uncorrectable error to be injected.
*/
static ssize_t i7core_inject_eccmask_store(struct device *dev,
struct device_attribute *mattr,
const char *data, size_t count)
{
struct mem_ctl_info *mci = to_mci(dev);
struct i7core_pvt *pvt = mci->pvt_info;
unsigned long value;
int rc;
if (pvt->inject.enable)
disable_inject(mci);
rc = kstrtoul(data, 10, &value);
if (rc < 0)
return -EIO;
pvt->inject.eccmask = (u32) value;
return count;
}
static ssize_t i7core_inject_eccmask_show(struct device *dev,
struct device_attribute *mattr,
char *data)
{
struct mem_ctl_info *mci = to_mci(dev);
struct i7core_pvt *pvt = mci->pvt_info;
return sprintf(data, "0x%08x\n", pvt->inject.eccmask);
}
/*
* i7core_addrmatch
*
* The type of error (UE/CE) will depend on the inject.eccmask value:
* Any bits set to a 1 will flip the corresponding ECC bit
* Correctable errors can be injected by flipping 1 bit or the bits within
* a symbol pair (2 consecutive aligned 8-bit pairs - i.e. 7:0 and 15:8 or
* 23:16 and 31:24). Flipping bits in two symbol pairs will cause an
* uncorrectable error to be injected.
*/
#define DECLARE_ADDR_MATCH(param, limit) \
static ssize_t i7core_inject_store_##param( \
struct device *dev, \
struct device_attribute *mattr, \
const char *data, size_t count) \
{ \
struct mem_ctl_info *mci = dev_get_drvdata(dev); \
struct i7core_pvt *pvt; \
long value; \
int rc; \
\
edac_dbg(1, "\n"); \
pvt = mci->pvt_info; \
\
if (pvt->inject.enable) \
disable_inject(mci); \
\
if (!strcasecmp(data, "any") || !strcasecmp(data, "any\n"))\
value = -1; \
else { \
rc = kstrtoul(data, 10, &value); \
if ((rc < 0) || (value >= limit)) \
return -EIO; \
} \
\
pvt->inject.param = value; \
\
return count; \
} \
\
static ssize_t i7core_inject_show_##param( \
struct device *dev, \
struct device_attribute *mattr, \
char *data) \
{ \
struct mem_ctl_info *mci = dev_get_drvdata(dev); \
struct i7core_pvt *pvt; \
\
pvt = mci->pvt_info; \
edac_dbg(1, "pvt=%p\n", pvt); \
if (pvt->inject.param < 0) \
return sprintf(data, "any\n"); \
else \
return sprintf(data, "%d\n", pvt->inject.param);\
}
#define ATTR_ADDR_MATCH(param) \
static DEVICE_ATTR(param, S_IRUGO | S_IWUSR, \
i7core_inject_show_##param, \
i7core_inject_store_##param)
DECLARE_ADDR_MATCH(channel, 3);
DECLARE_ADDR_MATCH(dimm, 3);
DECLARE_ADDR_MATCH(rank, 4);
DECLARE_ADDR_MATCH(bank, 32);
DECLARE_ADDR_MATCH(page, 0x10000);
DECLARE_ADDR_MATCH(col, 0x4000);
ATTR_ADDR_MATCH(channel);
ATTR_ADDR_MATCH(dimm);
ATTR_ADDR_MATCH(rank);
ATTR_ADDR_MATCH(bank);
ATTR_ADDR_MATCH(page);
ATTR_ADDR_MATCH(col);
static int write_and_test(struct pci_dev *dev, const int where, const u32 val)
{
u32 read;
int count;
edac_dbg(0, "setting pci %02x:%02x.%x reg=%02x value=%08x\n",
dev->bus->number, PCI_SLOT(dev->devfn), PCI_FUNC(dev->devfn),
where, val);
for (count = 0; count < 10; count++) {
if (count)
msleep(100);
pci_write_config_dword(dev, where, val);
pci_read_config_dword(dev, where, &read);
if (read == val)
return 0;
}
i7core_printk(KERN_ERR, "Error during set pci %02x:%02x.%x reg=%02x "
"write=%08x. Read=%08x\n",
dev->bus->number, PCI_SLOT(dev->devfn), PCI_FUNC(dev->devfn),
where, val, read);
return -EINVAL;
}
/*
* This routine prepares the Memory Controller for error injection.
* The error will be injected when some process tries to write to the
* memory that matches the given criteria.
* The criteria can be set in terms of a mask where dimm, rank, bank, page
* and col can be specified.
* A -1 value for any of the mask items will make the MCU to ignore
* that matching criteria for error injection.
*
* It should be noticed that the error will only happen after a write operation
* on a memory that matches the condition. if REPEAT_EN is not enabled at
* inject mask, then it will produce just one error. Otherwise, it will repeat
* until the injectmask would be cleaned.
*
* FIXME: This routine assumes that MAXNUMDIMMS value of MC_MAX_DOD
* is reliable enough to check if the MC is using the
* three channels. However, this is not clear at the datasheet.
*/
static ssize_t i7core_inject_enable_store(struct device *dev,
struct device_attribute *mattr,
const char *data, size_t count)
{
struct mem_ctl_info *mci = to_mci(dev);
struct i7core_pvt *pvt = mci->pvt_info;
u32 injectmask;
u64 mask = 0;
int rc;
long enable;
if (!pvt->pci_ch[pvt->inject.channel][0])
return 0;
rc = kstrtoul(data, 10, &enable);
if ((rc < 0))
return 0;
if (enable) {
pvt->inject.enable = 1;
} else {
disable_inject(mci);
return count;
}
/* Sets pvt->inject.dimm mask */
if (pvt->inject.dimm < 0)
mask |= 1LL << 41;
else {
if (pvt->channel[pvt->inject.channel].dimms > 2)
mask |= (pvt->inject.dimm & 0x3LL) << 35;
else
mask |= (pvt->inject.dimm & 0x1LL) << 36;
}
/* Sets pvt->inject.rank mask */
if (pvt->inject.rank < 0)
mask |= 1LL << 40;
else {
if (pvt->channel[pvt->inject.channel].dimms > 2)
mask |= (pvt->inject.rank & 0x1LL) << 34;
else
mask |= (pvt->inject.rank & 0x3LL) << 34;
}
/* Sets pvt->inject.bank mask */
if (pvt->inject.bank < 0)
mask |= 1LL << 39;
else
mask |= (pvt->inject.bank & 0x15LL) << 30;
/* Sets pvt->inject.page mask */
if (pvt->inject.page < 0)
mask |= 1LL << 38;
else
mask |= (pvt->inject.page & 0xffff) << 14;
/* Sets pvt->inject.column mask */
if (pvt->inject.col < 0)
mask |= 1LL << 37;
else
mask |= (pvt->inject.col & 0x3fff);
/*
* bit 0: REPEAT_EN
* bits 1-2: MASK_HALF_CACHELINE
* bit 3: INJECT_ECC
* bit 4: INJECT_ADDR_PARITY
*/
injectmask = (pvt->inject.type & 1) |
(pvt->inject.section & 0x3) << 1 |
(pvt->inject.type & 0x6) << (3 - 1);
/* Unlock writes to registers - this register is write only */
pci_write_config_dword(pvt->pci_noncore,
MC_CFG_CONTROL, 0x2);
write_and_test(pvt->pci_ch[pvt->inject.channel][0],
MC_CHANNEL_ADDR_MATCH, mask);
write_and_test(pvt->pci_ch[pvt->inject.channel][0],
MC_CHANNEL_ADDR_MATCH + 4, mask >> 32L);
write_and_test(pvt->pci_ch[pvt->inject.channel][0],
MC_CHANNEL_ERROR_MASK, pvt->inject.eccmask);
write_and_test(pvt->pci_ch[pvt->inject.channel][0],
MC_CHANNEL_ERROR_INJECT, injectmask);
/*
* This is something undocumented, based on my tests
* Without writing 8 to this register, errors aren't injected. Not sure
* why.
*/
pci_write_config_dword(pvt->pci_noncore,
MC_CFG_CONTROL, 8);
edac_dbg(0, "Error inject addr match 0x%016llx, ecc 0x%08x, inject 0x%08x\n",
mask, pvt->inject.eccmask, injectmask);
return count;
}
static ssize_t i7core_inject_enable_show(struct device *dev,
struct device_attribute *mattr,
char *data)
{
struct mem_ctl_info *mci = to_mci(dev);
struct i7core_pvt *pvt = mci->pvt_info;
u32 injectmask;
if (!pvt->pci_ch[pvt->inject.channel][0])
return 0;
pci_read_config_dword(pvt->pci_ch[pvt->inject.channel][0],
MC_CHANNEL_ERROR_INJECT, &injectmask);
edac_dbg(0, "Inject error read: 0x%018x\n", injectmask);
if (injectmask & 0x0c)
pvt->inject.enable = 1;
return sprintf(data, "%d\n", pvt->inject.enable);
}
#define DECLARE_COUNTER(param) \
static ssize_t i7core_show_counter_##param( \
struct device *dev, \
struct device_attribute *mattr, \
char *data) \
{ \
struct mem_ctl_info *mci = dev_get_drvdata(dev); \
struct i7core_pvt *pvt = mci->pvt_info; \
\
edac_dbg(1, "\n"); \
if (!pvt->ce_count_available || (pvt->is_registered)) \
return sprintf(data, "data unavailable\n"); \
return sprintf(data, "%lu\n", \
pvt->udimm_ce_count[param]); \
}
#define ATTR_COUNTER(param) \
static DEVICE_ATTR(udimm##param, S_IRUGO | S_IWUSR, \
i7core_show_counter_##param, \
NULL)
DECLARE_COUNTER(0);
DECLARE_COUNTER(1);
DECLARE_COUNTER(2);
ATTR_COUNTER(0);
ATTR_COUNTER(1);
ATTR_COUNTER(2);
/*
* inject_addrmatch device sysfs struct
*/
static struct attribute *i7core_addrmatch_attrs[] = {
&dev_attr_channel.attr,
&dev_attr_dimm.attr,
&dev_attr_rank.attr,
&dev_attr_bank.attr,
&dev_attr_page.attr,
&dev_attr_col.attr,
NULL
};
static struct attribute_group addrmatch_grp = {
.attrs = i7core_addrmatch_attrs,
};
static const struct attribute_group *addrmatch_groups[] = {
&addrmatch_grp,
NULL
};
static void addrmatch_release(struct device *device)
{
edac_dbg(1, "Releasing device %s\n", dev_name(device));
kfree(device);
}
static struct device_type addrmatch_type = {
.groups = addrmatch_groups,
.release = addrmatch_release,
};
/*
* all_channel_counts sysfs struct
*/
static struct attribute *i7core_udimm_counters_attrs[] = {
&dev_attr_udimm0.attr,
&dev_attr_udimm1.attr,
&dev_attr_udimm2.attr,
NULL
};
static struct attribute_group all_channel_counts_grp = {
.attrs = i7core_udimm_counters_attrs,
};
static const struct attribute_group *all_channel_counts_groups[] = {
&all_channel_counts_grp,
NULL
};
static void all_channel_counts_release(struct device *device)
{
edac_dbg(1, "Releasing device %s\n", dev_name(device));
kfree(device);
}
static struct device_type all_channel_counts_type = {
.groups = all_channel_counts_groups,
.release = all_channel_counts_release,
};
/*
* inject sysfs attributes
*/
static DEVICE_ATTR(inject_section, S_IRUGO | S_IWUSR,
i7core_inject_section_show, i7core_inject_section_store);
static DEVICE_ATTR(inject_type, S_IRUGO | S_IWUSR,
i7core_inject_type_show, i7core_inject_type_store);
static DEVICE_ATTR(inject_eccmask, S_IRUGO | S_IWUSR,
i7core_inject_eccmask_show, i7core_inject_eccmask_store);
static DEVICE_ATTR(inject_enable, S_IRUGO | S_IWUSR,
i7core_inject_enable_show, i7core_inject_enable_store);
static int i7core_create_sysfs_devices(struct mem_ctl_info *mci)
{
struct i7core_pvt *pvt = mci->pvt_info;
int rc;
rc = device_create_file(&mci->dev, &dev_attr_inject_section);
if (rc < 0)
return rc;
rc = device_create_file(&mci->dev, &dev_attr_inject_type);
if (rc < 0)
return rc;
rc = device_create_file(&mci->dev, &dev_attr_inject_eccmask);
if (rc < 0)
return rc;
rc = device_create_file(&mci->dev, &dev_attr_inject_enable);
if (rc < 0)
return rc;
pvt->addrmatch_dev = kzalloc(sizeof(*pvt->addrmatch_dev), GFP_KERNEL);
if (!pvt->addrmatch_dev)
return rc;
pvt->addrmatch_dev->type = &addrmatch_type;
pvt->addrmatch_dev->bus = mci->dev.bus;
device_initialize(pvt->addrmatch_dev);
pvt->addrmatch_dev->parent = &mci->dev;
dev_set_name(pvt->addrmatch_dev, "inject_addrmatch");
dev_set_drvdata(pvt->addrmatch_dev, mci);
edac_dbg(1, "creating %s\n", dev_name(pvt->addrmatch_dev));
rc = device_add(pvt->addrmatch_dev);
if (rc < 0)
return rc;
if (!pvt->is_registered) {
pvt->chancounts_dev = kzalloc(sizeof(*pvt->chancounts_dev),
GFP_KERNEL);
if (!pvt->chancounts_dev) {
put_device(pvt->addrmatch_dev);
device_del(pvt->addrmatch_dev);
return rc;
}
pvt->chancounts_dev->type = &all_channel_counts_type;
pvt->chancounts_dev->bus = mci->dev.bus;
device_initialize(pvt->chancounts_dev);
pvt->chancounts_dev->parent = &mci->dev;
dev_set_name(pvt->chancounts_dev, "all_channel_counts");
dev_set_drvdata(pvt->chancounts_dev, mci);
edac_dbg(1, "creating %s\n", dev_name(pvt->chancounts_dev));
rc = device_add(pvt->chancounts_dev);
if (rc < 0)
return rc;
}
return 0;
}
static void i7core_delete_sysfs_devices(struct mem_ctl_info *mci)
{
struct i7core_pvt *pvt = mci->pvt_info;
edac_dbg(1, "\n");
device_remove_file(&mci->dev, &dev_attr_inject_section);
device_remove_file(&mci->dev, &dev_attr_inject_type);
device_remove_file(&mci->dev, &dev_attr_inject_eccmask);
device_remove_file(&mci->dev, &dev_attr_inject_enable);
if (!pvt->is_registered) {
put_device(pvt->chancounts_dev);
device_del(pvt->chancounts_dev);
}
put_device(pvt->addrmatch_dev);
device_del(pvt->addrmatch_dev);
}
/****************************************************************************
Device initialization routines: put/get, init/exit
****************************************************************************/
/*
* i7core_put_all_devices 'put' all the devices that we have
* reserved via 'get'
*/
static void i7core_put_devices(struct i7core_dev *i7core_dev)
{
int i;
edac_dbg(0, "\n");
for (i = 0; i < i7core_dev->n_devs; i++) {
struct pci_dev *pdev = i7core_dev->pdev[i];
if (!pdev)
continue;
edac_dbg(0, "Removing dev %02x:%02x.%d\n",
pdev->bus->number,
PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn));
pci_dev_put(pdev);
}
}
static void i7core_put_all_devices(void)
{
struct i7core_dev *i7core_dev, *tmp;
list_for_each_entry_safe(i7core_dev, tmp, &i7core_edac_list, list) {
i7core_put_devices(i7core_dev);
free_i7core_dev(i7core_dev);
}
}
static void __init i7core_xeon_pci_fixup(const struct pci_id_table *table)
{
struct pci_dev *pdev = NULL;
int i;
/*
* On Xeon 55xx, the Intel Quick Path Arch Generic Non-core pci buses
* aren't announced by acpi. So, we need to use a legacy scan probing
* to detect them
*/
while (table && table->descr) {
pdev = pci_get_device(PCI_VENDOR_ID_INTEL, table->descr[0].dev_id, NULL);
if (unlikely(!pdev)) {
for (i = 0; i < MAX_SOCKET_BUSES; i++)
pcibios_scan_specific_bus(255-i);
}
pci_dev_put(pdev);
table++;
}
}
static unsigned i7core_pci_lastbus(void)
{
int last_bus = 0, bus;
struct pci_bus *b = NULL;
while ((b = pci_find_next_bus(b)) != NULL) {
bus = b->number;
edac_dbg(0, "Found bus %d\n", bus);
if (bus > last_bus)
last_bus = bus;
}
edac_dbg(0, "Last bus %d\n", last_bus);
return last_bus;
}
/*
* i7core_get_all_devices Find and perform 'get' operation on the MCH's
* device/functions we want to reference for this driver
*
* Need to 'get' device 16 func 1 and func 2
*/
static int i7core_get_onedevice(struct pci_dev **prev,
const struct pci_id_table *table,
const unsigned devno,
const unsigned last_bus)
{
struct i7core_dev *i7core_dev;
const struct pci_id_descr *dev_descr = &table->descr[devno];
struct pci_dev *pdev = NULL;
u8 bus = 0;
u8 socket = 0;
pdev = pci_get_device(PCI_VENDOR_ID_INTEL,
dev_descr->dev_id, *prev);
/*
* On Xeon 55xx, the Intel QuickPath Arch Generic Non-core regs
* is at addr 8086:2c40, instead of 8086:2c41. So, we need
* to probe for the alternate address in case of failure
*/
if (dev_descr->dev_id == PCI_DEVICE_ID_INTEL_I7_NONCORE && !pdev)
pdev = pci_get_device(PCI_VENDOR_ID_INTEL,
PCI_DEVICE_ID_INTEL_I7_NONCORE_ALT, *prev);
if (dev_descr->dev_id == PCI_DEVICE_ID_INTEL_LYNNFIELD_NONCORE && !pdev)
pdev = pci_get_device(PCI_VENDOR_ID_INTEL,
PCI_DEVICE_ID_INTEL_LYNNFIELD_NONCORE_ALT,
*prev);
if (!pdev) {
if (*prev) {
*prev = pdev;
return 0;
}
if (dev_descr->optional)
return 0;
if (devno == 0)
return -ENODEV;
i7core_printk(KERN_INFO,
"Device not found: dev %02x.%d PCI ID %04x:%04x\n",
dev_descr->dev, dev_descr->func,
PCI_VENDOR_ID_INTEL, dev_descr->dev_id);
/* End of list, leave */
return -ENODEV;
}
bus = pdev->bus->number;
socket = last_bus - bus;
i7core_dev = get_i7core_dev(socket);
if (!i7core_dev) {
i7core_dev = alloc_i7core_dev(socket, table);
if (!i7core_dev) {
pci_dev_put(pdev);
return -ENOMEM;
}
}
if (i7core_dev->pdev[devno]) {
i7core_printk(KERN_ERR,
"Duplicated device for "
"dev %02x:%02x.%d PCI ID %04x:%04x\n",
bus, dev_descr->dev, dev_descr->func,
PCI_VENDOR_ID_INTEL, dev_descr->dev_id);
pci_dev_put(pdev);
return -ENODEV;
}
i7core_dev->pdev[devno] = pdev;
/* Sanity check */
if (unlikely(PCI_SLOT(pdev->devfn) != dev_descr->dev ||
PCI_FUNC(pdev->devfn) != dev_descr->func)) {
i7core_printk(KERN_ERR,
"Device PCI ID %04x:%04x "
"has dev %02x:%02x.%d instead of dev %02x:%02x.%d\n",
PCI_VENDOR_ID_INTEL, dev_descr->dev_id,
bus, PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn),
bus, dev_descr->dev, dev_descr->func);
return -ENODEV;
}
/* Be sure that the device is enabled */
if (unlikely(pci_enable_device(pdev) < 0)) {
i7core_printk(KERN_ERR,
"Couldn't enable "
"dev %02x:%02x.%d PCI ID %04x:%04x\n",
bus, dev_descr->dev, dev_descr->func,
PCI_VENDOR_ID_INTEL, dev_descr->dev_id);
return -ENODEV;
}
edac_dbg(0, "Detected socket %d dev %02x:%02x.%d PCI ID %04x:%04x\n",
socket, bus, dev_descr->dev,
dev_descr->func,
PCI_VENDOR_ID_INTEL, dev_descr->dev_id);
/*
* As stated on drivers/pci/search.c, the reference count for
* @from is always decremented if it is not %NULL. So, as we need
* to get all devices up to null, we need to do a get for the device
*/
pci_dev_get(pdev);
*prev = pdev;
return 0;
}
static int i7core_get_all_devices(void)
{
int i, rc, last_bus;
struct pci_dev *pdev = NULL;
const struct pci_id_table *table = pci_dev_table;
last_bus = i7core_pci_lastbus();
while (table && table->descr) {
for (i = 0; i < table->n_devs; i++) {
pdev = NULL;
do {
rc = i7core_get_onedevice(&pdev, table, i,
last_bus);
if (rc < 0) {
if (i == 0) {
i = table->n_devs;
break;
}
i7core_put_all_devices();
return -ENODEV;
}
} while (pdev);
}
table++;
}
return 0;
}
static int mci_bind_devs(struct mem_ctl_info *mci,
struct i7core_dev *i7core_dev)
{
struct i7core_pvt *pvt = mci->pvt_info;
struct pci_dev *pdev;
int i, func, slot;
char *family;
pvt->is_registered = false;
pvt->enable_scrub = false;
for (i = 0; i < i7core_dev->n_devs; i++) {
pdev = i7core_dev->pdev[i];
if (!pdev)
continue;
func = PCI_FUNC(pdev->devfn);
slot = PCI_SLOT(pdev->devfn);
if (slot == 3) {
if (unlikely(func > MAX_MCR_FUNC))
goto error;
pvt->pci_mcr[func] = pdev;
} else if (likely(slot >= 4 && slot < 4 + NUM_CHANS)) {
if (unlikely(func > MAX_CHAN_FUNC))
goto error;
pvt->pci_ch[slot - 4][func] = pdev;
} else if (!slot && !func) {
pvt->pci_noncore = pdev;
/* Detect the processor family */
switch (pdev->device) {
case PCI_DEVICE_ID_INTEL_I7_NONCORE:
family = "Xeon 35xx/ i7core";
pvt->enable_scrub = false;
break;
case PCI_DEVICE_ID_INTEL_LYNNFIELD_NONCORE_ALT:
family = "i7-800/i5-700";
pvt->enable_scrub = false;
break;
case PCI_DEVICE_ID_INTEL_LYNNFIELD_NONCORE:
family = "Xeon 34xx";
pvt->enable_scrub = false;
break;
case PCI_DEVICE_ID_INTEL_I7_NONCORE_ALT:
family = "Xeon 55xx";
pvt->enable_scrub = true;
break;
case PCI_DEVICE_ID_INTEL_LYNNFIELD_NONCORE_REV2:
family = "Xeon 56xx / i7-900";
pvt->enable_scrub = true;
break;
default:
family = "unknown";
pvt->enable_scrub = false;
}
edac_dbg(0, "Detected a processor type %s\n", family);
} else
goto error;
edac_dbg(0, "Associated fn %d.%d, dev = %p, socket %d\n",
PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn),
pdev, i7core_dev->socket);
if (PCI_SLOT(pdev->devfn) == 3 &&
PCI_FUNC(pdev->devfn) == 2)
pvt->is_registered = true;
}
return 0;
error:
i7core_printk(KERN_ERR, "Device %d, function %d "
"is out of the expected range\n",
slot, func);
return -EINVAL;
}
/****************************************************************************
Error check routines
****************************************************************************/
static void i7core_rdimm_update_ce_count(struct mem_ctl_info *mci,
const int chan,
const int new0,
const int new1,
const int new2)
{
struct i7core_pvt *pvt = mci->pvt_info;
int add0 = 0, add1 = 0, add2 = 0;
/* Updates CE counters if it is not the first time here */
if (pvt->ce_count_available) {
/* Updates CE counters */
add2 = new2 - pvt->rdimm_last_ce_count[chan][2];
add1 = new1 - pvt->rdimm_last_ce_count[chan][1];
add0 = new0 - pvt->rdimm_last_ce_count[chan][0];
if (add2 < 0)
add2 += 0x7fff;
pvt->rdimm_ce_count[chan][2] += add2;
if (add1 < 0)
add1 += 0x7fff;
pvt->rdimm_ce_count[chan][1] += add1;
if (add0 < 0)
add0 += 0x7fff;
pvt->rdimm_ce_count[chan][0] += add0;
} else
pvt->ce_count_available = 1;
/* Store the new values */
pvt->rdimm_last_ce_count[chan][2] = new2;
pvt->rdimm_last_ce_count[chan][1] = new1;
pvt->rdimm_last_ce_count[chan][0] = new0;
/*updated the edac core */
if (add0 != 0)
edac_mc_handle_error(HW_EVENT_ERR_CORRECTED, mci, add0,
0, 0, 0,
chan, 0, -1, "error", "");
if (add1 != 0)
edac_mc_handle_error(HW_EVENT_ERR_CORRECTED, mci, add1,
0, 0, 0,
chan, 1, -1, "error", "");
if (add2 != 0)
edac_mc_handle_error(HW_EVENT_ERR_CORRECTED, mci, add2,
0, 0, 0,
chan, 2, -1, "error", "");
}
static void i7core_rdimm_check_mc_ecc_err(struct mem_ctl_info *mci)
{
struct i7core_pvt *pvt = mci->pvt_info;
u32 rcv[3][2];
int i, new0, new1, new2;
/*Read DEV 3: FUN 2: MC_COR_ECC_CNT regs directly*/
pci_read_config_dword(pvt->pci_mcr[2], MC_COR_ECC_CNT_0,
&rcv[0][0]);
pci_read_config_dword(pvt->pci_mcr[2], MC_COR_ECC_CNT_1,
&rcv[0][1]);
pci_read_config_dword(pvt->pci_mcr[2], MC_COR_ECC_CNT_2,
&rcv[1][0]);
pci_read_config_dword(pvt->pci_mcr[2], MC_COR_ECC_CNT_3,
&rcv[1][1]);
pci_read_config_dword(pvt->pci_mcr[2], MC_COR_ECC_CNT_4,
&rcv[2][0]);
pci_read_config_dword(pvt->pci_mcr[2], MC_COR_ECC_CNT_5,
&rcv[2][1]);
for (i = 0 ; i < 3; i++) {
edac_dbg(3, "MC_COR_ECC_CNT%d = 0x%x; MC_COR_ECC_CNT%d = 0x%x\n",
(i * 2), rcv[i][0], (i * 2) + 1, rcv[i][1]);
/*if the channel has 3 dimms*/
if (pvt->channel[i].dimms > 2) {
new0 = DIMM_BOT_COR_ERR(rcv[i][0]);
new1 = DIMM_TOP_COR_ERR(rcv[i][0]);
new2 = DIMM_BOT_COR_ERR(rcv[i][1]);
} else {
new0 = DIMM_TOP_COR_ERR(rcv[i][0]) +
DIMM_BOT_COR_ERR(rcv[i][0]);
new1 = DIMM_TOP_COR_ERR(rcv[i][1]) +
DIMM_BOT_COR_ERR(rcv[i][1]);
new2 = 0;
}
i7core_rdimm_update_ce_count(mci, i, new0, new1, new2);
}
}
/* This function is based on the device 3 function 4 registers as described on:
* Intel Xeon Processor 5500 Series Datasheet Volume 2
* http://www.intel.com/Assets/PDF/datasheet/321322.pdf
* also available at:
* http://www.arrownac.com/manufacturers/intel/s/nehalem/5500-datasheet-v2.pdf
*/
static void i7core_udimm_check_mc_ecc_err(struct mem_ctl_info *mci)
{
struct i7core_pvt *pvt = mci->pvt_info;
u32 rcv1, rcv0;
int new0, new1, new2;
if (!pvt->pci_mcr[4]) {
edac_dbg(0, "MCR registers not found\n");
return;
}
/* Corrected test errors */
pci_read_config_dword(pvt->pci_mcr[4], MC_TEST_ERR_RCV1, &rcv1);
pci_read_config_dword(pvt->pci_mcr[4], MC_TEST_ERR_RCV0, &rcv0);
/* Store the new values */
new2 = DIMM2_COR_ERR(rcv1);
new1 = DIMM1_COR_ERR(rcv0);
new0 = DIMM0_COR_ERR(rcv0);
/* Updates CE counters if it is not the first time here */
if (pvt->ce_count_available) {
/* Updates CE counters */
int add0, add1, add2;
add2 = new2 - pvt->udimm_last_ce_count[2];
add1 = new1 - pvt->udimm_last_ce_count[1];
add0 = new0 - pvt->udimm_last_ce_count[0];
if (add2 < 0)
add2 += 0x7fff;
pvt->udimm_ce_count[2] += add2;
if (add1 < 0)
add1 += 0x7fff;
pvt->udimm_ce_count[1] += add1;
if (add0 < 0)
add0 += 0x7fff;
pvt->udimm_ce_count[0] += add0;
if (add0 | add1 | add2)
i7core_printk(KERN_ERR, "New Corrected error(s): "
"dimm0: +%d, dimm1: +%d, dimm2 +%d\n",
add0, add1, add2);
} else
pvt->ce_count_available = 1;
/* Store the new values */
pvt->udimm_last_ce_count[2] = new2;
pvt->udimm_last_ce_count[1] = new1;
pvt->udimm_last_ce_count[0] = new0;
}
/*
* According with tables E-11 and E-12 of chapter E.3.3 of Intel 64 and IA-32
* Architectures Software Developers Manual Volume 3B.
* Nehalem are defined as family 0x06, model 0x1a
*
* The MCA registers used here are the following ones:
* struct mce field MCA Register
* m->status MSR_IA32_MC8_STATUS
* m->addr MSR_IA32_MC8_ADDR
* m->misc MSR_IA32_MC8_MISC
* In the case of Nehalem, the error information is masked at .status and .misc
* fields
*/
static void i7core_mce_output_error(struct mem_ctl_info *mci,
const struct mce *m)
{
struct i7core_pvt *pvt = mci->pvt_info;
char *type, *optype, *err;
enum hw_event_mc_err_type tp_event;
unsigned long error = m->status & 0x1ff0000l;
bool uncorrected_error = m->mcgstatus & 1ll << 61;
bool ripv = m->mcgstatus & 1;
u32 optypenum = (m->status >> 4) & 0x07;
u32 core_err_cnt = (m->status >> 38) & 0x7fff;
u32 dimm = (m->misc >> 16) & 0x3;
u32 channel = (m->misc >> 18) & 0x3;
u32 syndrome = m->misc >> 32;
u32 errnum = find_first_bit(&error, 32);
if (uncorrected_error) {
if (ripv) {
type = "FATAL";
tp_event = HW_EVENT_ERR_FATAL;
} else {
type = "NON_FATAL";
tp_event = HW_EVENT_ERR_UNCORRECTED;
}
} else {
type = "CORRECTED";
tp_event = HW_EVENT_ERR_CORRECTED;
}
switch (optypenum) {
case 0:
optype = "generic undef request";
break;
case 1:
optype = "read error";
break;
case 2:
optype = "write error";
break;
case 3:
optype = "addr/cmd error";
break;
case 4:
optype = "scrubbing error";
break;
default:
optype = "reserved";
break;
}
switch (errnum) {
case 16:
err = "read ECC error";
break;
case 17:
err = "RAS ECC error";
break;
case 18:
err = "write parity error";
break;
case 19:
err = "redundacy loss";
break;
case 20:
err = "reserved";
break;
case 21:
err = "memory range error";
break;
case 22:
err = "RTID out of range";
break;
case 23:
err = "address parity error";
break;
case 24:
err = "byte enable parity error";
break;
default:
err = "unknown";
}
/*
* Call the helper to output message
* FIXME: what to do if core_err_cnt > 1? Currently, it generates
* only one event
*/
if (uncorrected_error || !pvt->is_registered)
edac_mc_handle_error(tp_event, mci, core_err_cnt,
m->addr >> PAGE_SHIFT,
m->addr & ~PAGE_MASK,
syndrome,
channel, dimm, -1,
err, optype);
}
/*
* i7core_check_error Retrieve and process errors reported by the
* hardware. Called by the Core module.
*/
static void i7core_check_error(struct mem_ctl_info *mci)
{
struct i7core_pvt *pvt = mci->pvt_info;
int i;
unsigned count = 0;
struct mce *m;
/*
* MCE first step: Copy all mce errors into a temporary buffer
* We use a double buffering here, to reduce the risk of
* losing an error.
*/
smp_rmb();
count = (pvt->mce_out + MCE_LOG_LEN - pvt->mce_in)
% MCE_LOG_LEN;
if (!count)
goto check_ce_error;
m = pvt->mce_outentry;
if (pvt->mce_in + count > MCE_LOG_LEN) {
unsigned l = MCE_LOG_LEN - pvt->mce_in;
memcpy(m, &pvt->mce_entry[pvt->mce_in], sizeof(*m) * l);
smp_wmb();
pvt->mce_in = 0;
count -= l;
m += l;
}
memcpy(m, &pvt->mce_entry[pvt->mce_in], sizeof(*m) * count);
smp_wmb();
pvt->mce_in += count;
smp_rmb();
if (pvt->mce_overrun) {
i7core_printk(KERN_ERR, "Lost %d memory errors\n",
pvt->mce_overrun);
smp_wmb();
pvt->mce_overrun = 0;
}
/*
* MCE second step: parse errors and display
*/
for (i = 0; i < count; i++)
i7core_mce_output_error(mci, &pvt->mce_outentry[i]);
/*
* Now, let's increment CE error counts
*/
check_ce_error:
if (!pvt->is_registered)
i7core_udimm_check_mc_ecc_err(mci);
else
i7core_rdimm_check_mc_ecc_err(mci);
}
/*
* i7core_mce_check_error Replicates mcelog routine to get errors
* This routine simply queues mcelog errors, and
* return. The error itself should be handled later
* by i7core_check_error.
* WARNING: As this routine should be called at NMI time, extra care should
* be taken to avoid deadlocks, and to be as fast as possible.
*/
static int i7core_mce_check_error(struct notifier_block *nb, unsigned long val,
void *data)
{
struct mce *mce = (struct mce *)data;
struct i7core_dev *i7_dev;
struct mem_ctl_info *mci;
struct i7core_pvt *pvt;
i7_dev = get_i7core_dev(mce->socketid);
if (!i7_dev)
return NOTIFY_BAD;
mci = i7_dev->mci;
pvt = mci->pvt_info;
/*
* Just let mcelog handle it if the error is
* outside the memory controller
*/
if (((mce->status & 0xffff) >> 7) != 1)
return NOTIFY_DONE;
/* Bank 8 registers are the only ones that we know how to handle */
if (mce->bank != 8)
return NOTIFY_DONE;
smp_rmb();
if ((pvt->mce_out + 1) % MCE_LOG_LEN == pvt->mce_in) {
smp_wmb();
pvt->mce_overrun++;
return NOTIFY_DONE;
}
/* Copy memory error at the ringbuffer */
memcpy(&pvt->mce_entry[pvt->mce_out], mce, sizeof(*mce));
smp_wmb();
pvt->mce_out = (pvt->mce_out + 1) % MCE_LOG_LEN;
/* Handle fatal errors immediately */
if (mce->mcgstatus & 1)
i7core_check_error(mci);
/* Advise mcelog that the errors were handled */
return NOTIFY_STOP;
}
static struct notifier_block i7_mce_dec = {
.notifier_call = i7core_mce_check_error,
};
struct memdev_dmi_entry {
u8 type;
u8 length;
u16 handle;
u16 phys_mem_array_handle;
u16 mem_err_info_handle;
u16 total_width;
u16 data_width;
u16 size;
u8 form;
u8 device_set;
u8 device_locator;
u8 bank_locator;
u8 memory_type;
u16 type_detail;
u16 speed;
u8 manufacturer;
u8 serial_number;
u8 asset_tag;
u8 part_number;
u8 attributes;
u32 extended_size;
u16 conf_mem_clk_speed;
} __attribute__((__packed__));
/*
* Decode the DRAM Clock Frequency, be paranoid, make sure that all
* memory devices show the same speed, and if they don't then consider
* all speeds to be invalid.
*/
static void decode_dclk(const struct dmi_header *dh, void *_dclk_freq)
{
int *dclk_freq = _dclk_freq;
u16 dmi_mem_clk_speed;
if (*dclk_freq == -1)
return;
if (dh->type == DMI_ENTRY_MEM_DEVICE) {
struct memdev_dmi_entry *memdev_dmi_entry =
(struct memdev_dmi_entry *)dh;
unsigned long conf_mem_clk_speed_offset =
(unsigned long)&memdev_dmi_entry->conf_mem_clk_speed -
(unsigned long)&memdev_dmi_entry->type;
unsigned long speed_offset =
(unsigned long)&memdev_dmi_entry->speed -
(unsigned long)&memdev_dmi_entry->type;
/* Check that a DIMM is present */
if (memdev_dmi_entry->size == 0)
return;
/*
* Pick the configured speed if it's available, otherwise
* pick the DIMM speed, or we don't have a speed.
*/
if (memdev_dmi_entry->length > conf_mem_clk_speed_offset) {
dmi_mem_clk_speed =
memdev_dmi_entry->conf_mem_clk_speed;
} else if (memdev_dmi_entry->length > speed_offset) {
dmi_mem_clk_speed = memdev_dmi_entry->speed;
} else {
*dclk_freq = -1;
return;
}
if (*dclk_freq == 0) {
/* First pass, speed was 0 */
if (dmi_mem_clk_speed > 0) {
/* Set speed if a valid speed is read */
*dclk_freq = dmi_mem_clk_speed;
} else {
/* Otherwise we don't have a valid speed */
*dclk_freq = -1;
}
} else if (*dclk_freq > 0 &&
*dclk_freq != dmi_mem_clk_speed) {
/*
* If we have a speed, check that all DIMMS are the same
* speed, otherwise set the speed as invalid.
*/
*dclk_freq = -1;
}
}
}
/*
* The default DCLK frequency is used as a fallback if we
* fail to find anything reliable in the DMI. The value
* is taken straight from the datasheet.
*/
#define DEFAULT_DCLK_FREQ 800
static int get_dclk_freq(void)
{
int dclk_freq = 0;
dmi_walk(decode_dclk, (void *)&dclk_freq);
if (dclk_freq < 1)
return DEFAULT_DCLK_FREQ;
return dclk_freq;
}
/*
* set_sdram_scrub_rate This routine sets byte/sec bandwidth scrub rate
* to hardware according to SCRUBINTERVAL formula
* found in datasheet.
*/
static int set_sdram_scrub_rate(struct mem_ctl_info *mci, u32 new_bw)
{
struct i7core_pvt *pvt = mci->pvt_info;
struct pci_dev *pdev;
u32 dw_scrub;
u32 dw_ssr;
/* Get data from the MC register, function 2 */
pdev = pvt->pci_mcr[2];
if (!pdev)
return -ENODEV;
pci_read_config_dword(pdev, MC_SCRUB_CONTROL, &dw_scrub);
if (new_bw == 0) {
/* Prepare to disable petrol scrub */
dw_scrub &= ~STARTSCRUB;
/* Stop the patrol scrub engine */
write_and_test(pdev, MC_SCRUB_CONTROL,
dw_scrub & ~SCRUBINTERVAL_MASK);
/* Get current status of scrub rate and set bit to disable */
pci_read_config_dword(pdev, MC_SSRCONTROL, &dw_ssr);
dw_ssr &= ~SSR_MODE_MASK;
dw_ssr |= SSR_MODE_DISABLE;
} else {
const int cache_line_size = 64;
const u32 freq_dclk_mhz = pvt->dclk_freq;
unsigned long long scrub_interval;
/*
* Translate the desired scrub rate to a register value and
* program the corresponding register value.
*/
scrub_interval = (unsigned long long)freq_dclk_mhz *
cache_line_size * 1000000;
do_div(scrub_interval, new_bw);
if (!scrub_interval || scrub_interval > SCRUBINTERVAL_MASK)
return -EINVAL;
dw_scrub = SCRUBINTERVAL_MASK & scrub_interval;
/* Start the patrol scrub engine */
pci_write_config_dword(pdev, MC_SCRUB_CONTROL,
STARTSCRUB | dw_scrub);
/* Get current status of scrub rate and set bit to enable */
pci_read_config_dword(pdev, MC_SSRCONTROL, &dw_ssr);
dw_ssr &= ~SSR_MODE_MASK;
dw_ssr |= SSR_MODE_ENABLE;
}
/* Disable or enable scrubbing */
pci_write_config_dword(pdev, MC_SSRCONTROL, dw_ssr);
return new_bw;
}
/*
* get_sdram_scrub_rate This routine convert current scrub rate value
* into byte/sec bandwidth according to
* SCRUBINTERVAL formula found in datasheet.
*/
static int get_sdram_scrub_rate(struct mem_ctl_info *mci)
{
struct i7core_pvt *pvt = mci->pvt_info;
struct pci_dev *pdev;
const u32 cache_line_size = 64;
const u32 freq_dclk_mhz = pvt->dclk_freq;
unsigned long long scrub_rate;
u32 scrubval;
/* Get data from the MC register, function 2 */
pdev = pvt->pci_mcr[2];
if (!pdev)
return -ENODEV;
/* Get current scrub control data */
pci_read_config_dword(pdev, MC_SCRUB_CONTROL, &scrubval);
/* Mask highest 8-bits to 0 */
scrubval &= SCRUBINTERVAL_MASK;
if (!scrubval)
return 0;
/* Calculate scrub rate value into byte/sec bandwidth */
scrub_rate = (unsigned long long)freq_dclk_mhz *
1000000 * cache_line_size;
do_div(scrub_rate, scrubval);
return (int)scrub_rate;
}
static void enable_sdram_scrub_setting(struct mem_ctl_info *mci)
{
struct i7core_pvt *pvt = mci->pvt_info;
u32 pci_lock;
/* Unlock writes to pci registers */
pci_read_config_dword(pvt->pci_noncore, MC_CFG_CONTROL, &pci_lock);
pci_lock &= ~0x3;
pci_write_config_dword(pvt->pci_noncore, MC_CFG_CONTROL,
pci_lock | MC_CFG_UNLOCK);
mci->set_sdram_scrub_rate = set_sdram_scrub_rate;
mci->get_sdram_scrub_rate = get_sdram_scrub_rate;
}
static void disable_sdram_scrub_setting(struct mem_ctl_info *mci)
{
struct i7core_pvt *pvt = mci->pvt_info;
u32 pci_lock;
/* Lock writes to pci registers */
pci_read_config_dword(pvt->pci_noncore, MC_CFG_CONTROL, &pci_lock);
pci_lock &= ~0x3;
pci_write_config_dword(pvt->pci_noncore, MC_CFG_CONTROL,
pci_lock | MC_CFG_LOCK);
}
static void i7core_pci_ctl_create(struct i7core_pvt *pvt)
{
pvt->i7core_pci = edac_pci_create_generic_ctl(
&pvt->i7core_dev->pdev[0]->dev,
EDAC_MOD_STR);
if (unlikely(!pvt->i7core_pci))
i7core_printk(KERN_WARNING,
"Unable to setup PCI error report via EDAC\n");
}
static void i7core_pci_ctl_release(struct i7core_pvt *pvt)
{
if (likely(pvt->i7core_pci))
edac_pci_release_generic_ctl(pvt->i7core_pci);
else
i7core_printk(KERN_ERR,
"Couldn't find mem_ctl_info for socket %d\n",
pvt->i7core_dev->socket);
pvt->i7core_pci = NULL;
}
static void i7core_unregister_mci(struct i7core_dev *i7core_dev)
{
struct mem_ctl_info *mci = i7core_dev->mci;
struct i7core_pvt *pvt;
if (unlikely(!mci || !mci->pvt_info)) {
edac_dbg(0, "MC: dev = %p\n", &i7core_dev->pdev[0]->dev);
i7core_printk(KERN_ERR, "Couldn't find mci handler\n");
return;
}
pvt = mci->pvt_info;
edac_dbg(0, "MC: mci = %p, dev = %p\n", mci, &i7core_dev->pdev[0]->dev);
/* Disable scrubrate setting */
if (pvt->enable_scrub)
disable_sdram_scrub_setting(mci);
/* Disable EDAC polling */
i7core_pci_ctl_release(pvt);
/* Remove MC sysfs nodes */
i7core_delete_sysfs_devices(mci);
edac_mc_del_mc(mci->pdev);
edac_dbg(1, "%s: free mci struct\n", mci->ctl_name);
kfree(mci->ctl_name);
edac_mc_free(mci);
i7core_dev->mci = NULL;
}
static int i7core_register_mci(struct i7core_dev *i7core_dev)
{
struct mem_ctl_info *mci;
struct i7core_pvt *pvt;
int rc;
struct edac_mc_layer layers[2];
/* allocate a new MC control structure */
layers[0].type = EDAC_MC_LAYER_CHANNEL;
layers[0].size = NUM_CHANS;
layers[0].is_virt_csrow = false;
layers[1].type = EDAC_MC_LAYER_SLOT;
layers[1].size = MAX_DIMMS;
layers[1].is_virt_csrow = true;
mci = edac_mc_alloc(i7core_dev->socket, ARRAY_SIZE(layers), layers,
sizeof(*pvt));
if (unlikely(!mci))
return -ENOMEM;
edac_dbg(0, "MC: mci = %p, dev = %p\n", mci, &i7core_dev->pdev[0]->dev);
pvt = mci->pvt_info;
memset(pvt, 0, sizeof(*pvt));
/* Associates i7core_dev and mci for future usage */
pvt->i7core_dev = i7core_dev;
i7core_dev->mci = mci;
/*
* FIXME: how to handle RDDR3 at MCI level? It is possible to have
* Mixed RDDR3/UDDR3 with Nehalem, provided that they are on different
* memory channels
*/
mci->mtype_cap = MEM_FLAG_DDR3;
mci->edac_ctl_cap = EDAC_FLAG_NONE;
mci->edac_cap = EDAC_FLAG_NONE;
mci->mod_name = "i7core_edac.c";
mci->mod_ver = I7CORE_REVISION;
mci->ctl_name = kasprintf(GFP_KERNEL, "i7 core #%d",
i7core_dev->socket);
mci->dev_name = pci_name(i7core_dev->pdev[0]);
mci->ctl_page_to_phys = NULL;
/* Store pci devices at mci for faster access */
rc = mci_bind_devs(mci, i7core_dev);
if (unlikely(rc < 0))
goto fail0;
/* Get dimm basic config */
get_dimm_config(mci);
/* record ptr to the generic device */
mci->pdev = &i7core_dev->pdev[0]->dev;
/* Set the function pointer to an actual operation function */
mci->edac_check = i7core_check_error;
/* Enable scrubrate setting */
if (pvt->enable_scrub)
enable_sdram_scrub_setting(mci);
/* add this new MC control structure to EDAC's list of MCs */
if (unlikely(edac_mc_add_mc(mci))) {
edac_dbg(0, "MC: failed edac_mc_add_mc()\n");
/* FIXME: perhaps some code should go here that disables error
* reporting if we just enabled it
*/
rc = -EINVAL;
goto fail0;
}
if (i7core_create_sysfs_devices(mci)) {
edac_dbg(0, "MC: failed to create sysfs nodes\n");
edac_mc_del_mc(mci->pdev);
rc = -EINVAL;
goto fail0;
}
/* Default error mask is any memory */
pvt->inject.channel = 0;
pvt->inject.dimm = -1;
pvt->inject.rank = -1;
pvt->inject.bank = -1;
pvt->inject.page = -1;
pvt->inject.col = -1;
/* allocating generic PCI control info */
i7core_pci_ctl_create(pvt);
/* DCLK for scrub rate setting */
pvt->dclk_freq = get_dclk_freq();
return 0;
fail0:
kfree(mci->ctl_name);
edac_mc_free(mci);
i7core_dev->mci = NULL;
return rc;
}
/*
* i7core_probe Probe for ONE instance of device to see if it is
* present.
* return:
* 0 for FOUND a device
* < 0 for error code
*/
static int i7core_probe(struct pci_dev *pdev, const struct pci_device_id *id)
{
int rc, count = 0;
struct i7core_dev *i7core_dev;
/* get the pci devices we want to reserve for our use */
mutex_lock(&i7core_edac_lock);
/*
* All memory controllers are allocated at the first pass.
*/
if (unlikely(probed >= 1)) {
mutex_unlock(&i7core_edac_lock);
return -ENODEV;
}
probed++;
rc = i7core_get_all_devices();
if (unlikely(rc < 0))
goto fail0;
list_for_each_entry(i7core_dev, &i7core_edac_list, list) {
count++;
rc = i7core_register_mci(i7core_dev);
if (unlikely(rc < 0))
goto fail1;
}
/*
* Nehalem-EX uses a different memory controller. However, as the
* memory controller is not visible on some Nehalem/Nehalem-EP, we
* need to indirectly probe via a X58 PCI device. The same devices
* are found on (some) Nehalem-EX. So, on those machines, the
* probe routine needs to return -ENODEV, as the actual Memory
* Controller registers won't be detected.
*/
if (!count) {
rc = -ENODEV;
goto fail1;
}
i7core_printk(KERN_INFO,
"Driver loaded, %d memory controller(s) found.\n",
count);
mutex_unlock(&i7core_edac_lock);
return 0;
fail1:
list_for_each_entry(i7core_dev, &i7core_edac_list, list)
i7core_unregister_mci(i7core_dev);
i7core_put_all_devices();
fail0:
mutex_unlock(&i7core_edac_lock);
return rc;
}
/*
* i7core_remove destructor for one instance of device
*
*/
static void i7core_remove(struct pci_dev *pdev)
{
struct i7core_dev *i7core_dev;
edac_dbg(0, "\n");
/*
* we have a trouble here: pdev value for removal will be wrong, since
* it will point to the X58 register used to detect that the machine
* is a Nehalem or upper design. However, due to the way several PCI
* devices are grouped together to provide MC functionality, we need
* to use a different method for releasing the devices
*/
mutex_lock(&i7core_edac_lock);
if (unlikely(!probed)) {
mutex_unlock(&i7core_edac_lock);
return;
}
list_for_each_entry(i7core_dev, &i7core_edac_list, list)
i7core_unregister_mci(i7core_dev);
/* Release PCI resources */
i7core_put_all_devices();
probed--;
mutex_unlock(&i7core_edac_lock);
}
MODULE_DEVICE_TABLE(pci, i7core_pci_tbl);
/*
* i7core_driver pci_driver structure for this module
*
*/
static struct pci_driver i7core_driver = {
.name = "i7core_edac",
.probe = i7core_probe,
.remove = i7core_remove,
.id_table = i7core_pci_tbl,
};
/*
* i7core_init Module entry function
* Try to initialize this module for its devices
*/
static int __init i7core_init(void)
{
int pci_rc;
edac_dbg(2, "\n");
/* Ensure that the OPSTATE is set correctly for POLL or NMI */
opstate_init();
if (use_pci_fixup)
i7core_xeon_pci_fixup(pci_dev_table);
pci_rc = pci_register_driver(&i7core_driver);
edac: avoid mce decoding crash after edac driver unloaded Some edac drivers register themselves as mce decoders via notifier_chain. But in current notifier_chain implementation logic, it doesn't accept same notifier registered twice. If so, it will be wrong when adding/removing the element from the list. For example, on one SandyBridge platform, remove module sb_edac and then trigger one error, it will hit oops because it has no mce decoder registered but related notifier_chain still points to an invalid callback function. Here is an example: Call Trace: [<ffffffff8150ef6a>] atomic_notifier_call_chain+0x1a/0x20 [<ffffffff8102b936>] mce_log+0x46/0x180 [<ffffffff8102eaea>] apei_mce_report_mem_error+0x4a/0x60 [<ffffffff812e19d2>] ghes_do_proc+0x192/0x210 [<ffffffff812e2066>] ghes_proc+0x46/0x70 [<ffffffff812e20d8>] ghes_notify_sci+0x48/0x80 [<ffffffff8150ef05>] notifier_call_chain+0x55/0x80 [<ffffffff81076f1a>] __blocking_notifier_call_chain+0x5a/0x80 [<ffffffff812aea11>] ? acpi_os_wait_events_complete+0x23/0x23 [<ffffffff81076f56>] blocking_notifier_call_chain+0x16/0x20 [<ffffffff812ddc4d>] acpi_hed_notify+0x19/0x1b [<ffffffff812b16bd>] acpi_device_notify+0x19/0x1b [<ffffffff812beb38>] acpi_ev_notify_dispatch+0x67/0x7f [<ffffffff812aea3a>] acpi_os_execute_deferred+0x29/0x36 [<ffffffff81069dc2>] process_one_work+0x132/0x450 [<ffffffff8106bbcb>] worker_thread+0x17b/0x3c0 [<ffffffff8106ba50>] ? manage_workers+0x120/0x120 [<ffffffff81070aee>] kthread+0x9e/0xb0 [<ffffffff81514724>] kernel_thread_helper+0x4/0x10 [<ffffffff81070a50>] ? kthread_freezable_should_stop+0x70/0x70 [<ffffffff81514720>] ? gs_change+0x13/0x13 Code: f3 49 89 d4 45 85 ed 4d 89 c6 48 8b 0f 74 48 48 85 c9 75 17 eb 41 0f 1f 80 00 00 00 00 41 83 ed 01 4c 89 f9 74 22 4d 85 ff 74 1d <4c> 8b 79 08 4c 89 e2 48 89 de 48 89 cf ff 11 4d 85 f6 74 04 41 RIP [<ffffffff8150eef6>] notifier_call_chain+0x46/0x80 RSP <ffff88042868fb20> CR2: ffffffffa01af838 ---[ end trace 0100930068e73e6f ]--- BUG: unable to handle kernel paging request at fffffffffffffff8 IP: [<ffffffff810705b0>] kthread_data+0x10/0x20 PGD 1a0d067 PUD 1a0e067 PMD 0 Oops: 0000 [#2] SMP Only i7core_edac and sb_edac have such issues because they have more than one memory controller which means they have to register mce decoder many times. Cc: <stable@vger.kernel.org> # 3.2 and upper Signed-off-by: Chen Gong <gong.chen@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2012-05-09 06:40:12 +07:00
if (pci_rc >= 0) {
mce_register_decode_chain(&i7_mce_dec);
return 0;
edac: avoid mce decoding crash after edac driver unloaded Some edac drivers register themselves as mce decoders via notifier_chain. But in current notifier_chain implementation logic, it doesn't accept same notifier registered twice. If so, it will be wrong when adding/removing the element from the list. For example, on one SandyBridge platform, remove module sb_edac and then trigger one error, it will hit oops because it has no mce decoder registered but related notifier_chain still points to an invalid callback function. Here is an example: Call Trace: [<ffffffff8150ef6a>] atomic_notifier_call_chain+0x1a/0x20 [<ffffffff8102b936>] mce_log+0x46/0x180 [<ffffffff8102eaea>] apei_mce_report_mem_error+0x4a/0x60 [<ffffffff812e19d2>] ghes_do_proc+0x192/0x210 [<ffffffff812e2066>] ghes_proc+0x46/0x70 [<ffffffff812e20d8>] ghes_notify_sci+0x48/0x80 [<ffffffff8150ef05>] notifier_call_chain+0x55/0x80 [<ffffffff81076f1a>] __blocking_notifier_call_chain+0x5a/0x80 [<ffffffff812aea11>] ? acpi_os_wait_events_complete+0x23/0x23 [<ffffffff81076f56>] blocking_notifier_call_chain+0x16/0x20 [<ffffffff812ddc4d>] acpi_hed_notify+0x19/0x1b [<ffffffff812b16bd>] acpi_device_notify+0x19/0x1b [<ffffffff812beb38>] acpi_ev_notify_dispatch+0x67/0x7f [<ffffffff812aea3a>] acpi_os_execute_deferred+0x29/0x36 [<ffffffff81069dc2>] process_one_work+0x132/0x450 [<ffffffff8106bbcb>] worker_thread+0x17b/0x3c0 [<ffffffff8106ba50>] ? manage_workers+0x120/0x120 [<ffffffff81070aee>] kthread+0x9e/0xb0 [<ffffffff81514724>] kernel_thread_helper+0x4/0x10 [<ffffffff81070a50>] ? kthread_freezable_should_stop+0x70/0x70 [<ffffffff81514720>] ? gs_change+0x13/0x13 Code: f3 49 89 d4 45 85 ed 4d 89 c6 48 8b 0f 74 48 48 85 c9 75 17 eb 41 0f 1f 80 00 00 00 00 41 83 ed 01 4c 89 f9 74 22 4d 85 ff 74 1d <4c> 8b 79 08 4c 89 e2 48 89 de 48 89 cf ff 11 4d 85 f6 74 04 41 RIP [<ffffffff8150eef6>] notifier_call_chain+0x46/0x80 RSP <ffff88042868fb20> CR2: ffffffffa01af838 ---[ end trace 0100930068e73e6f ]--- BUG: unable to handle kernel paging request at fffffffffffffff8 IP: [<ffffffff810705b0>] kthread_data+0x10/0x20 PGD 1a0d067 PUD 1a0e067 PMD 0 Oops: 0000 [#2] SMP Only i7core_edac and sb_edac have such issues because they have more than one memory controller which means they have to register mce decoder many times. Cc: <stable@vger.kernel.org> # 3.2 and upper Signed-off-by: Chen Gong <gong.chen@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2012-05-09 06:40:12 +07:00
}
i7core_printk(KERN_ERR, "Failed to register device with error %d.\n",
pci_rc);
return pci_rc;
}
/*
* i7core_exit() Module exit function
* Unregister the driver
*/
static void __exit i7core_exit(void)
{
edac_dbg(2, "\n");
pci_unregister_driver(&i7core_driver);
edac: avoid mce decoding crash after edac driver unloaded Some edac drivers register themselves as mce decoders via notifier_chain. But in current notifier_chain implementation logic, it doesn't accept same notifier registered twice. If so, it will be wrong when adding/removing the element from the list. For example, on one SandyBridge platform, remove module sb_edac and then trigger one error, it will hit oops because it has no mce decoder registered but related notifier_chain still points to an invalid callback function. Here is an example: Call Trace: [<ffffffff8150ef6a>] atomic_notifier_call_chain+0x1a/0x20 [<ffffffff8102b936>] mce_log+0x46/0x180 [<ffffffff8102eaea>] apei_mce_report_mem_error+0x4a/0x60 [<ffffffff812e19d2>] ghes_do_proc+0x192/0x210 [<ffffffff812e2066>] ghes_proc+0x46/0x70 [<ffffffff812e20d8>] ghes_notify_sci+0x48/0x80 [<ffffffff8150ef05>] notifier_call_chain+0x55/0x80 [<ffffffff81076f1a>] __blocking_notifier_call_chain+0x5a/0x80 [<ffffffff812aea11>] ? acpi_os_wait_events_complete+0x23/0x23 [<ffffffff81076f56>] blocking_notifier_call_chain+0x16/0x20 [<ffffffff812ddc4d>] acpi_hed_notify+0x19/0x1b [<ffffffff812b16bd>] acpi_device_notify+0x19/0x1b [<ffffffff812beb38>] acpi_ev_notify_dispatch+0x67/0x7f [<ffffffff812aea3a>] acpi_os_execute_deferred+0x29/0x36 [<ffffffff81069dc2>] process_one_work+0x132/0x450 [<ffffffff8106bbcb>] worker_thread+0x17b/0x3c0 [<ffffffff8106ba50>] ? manage_workers+0x120/0x120 [<ffffffff81070aee>] kthread+0x9e/0xb0 [<ffffffff81514724>] kernel_thread_helper+0x4/0x10 [<ffffffff81070a50>] ? kthread_freezable_should_stop+0x70/0x70 [<ffffffff81514720>] ? gs_change+0x13/0x13 Code: f3 49 89 d4 45 85 ed 4d 89 c6 48 8b 0f 74 48 48 85 c9 75 17 eb 41 0f 1f 80 00 00 00 00 41 83 ed 01 4c 89 f9 74 22 4d 85 ff 74 1d <4c> 8b 79 08 4c 89 e2 48 89 de 48 89 cf ff 11 4d 85 f6 74 04 41 RIP [<ffffffff8150eef6>] notifier_call_chain+0x46/0x80 RSP <ffff88042868fb20> CR2: ffffffffa01af838 ---[ end trace 0100930068e73e6f ]--- BUG: unable to handle kernel paging request at fffffffffffffff8 IP: [<ffffffff810705b0>] kthread_data+0x10/0x20 PGD 1a0d067 PUD 1a0e067 PMD 0 Oops: 0000 [#2] SMP Only i7core_edac and sb_edac have such issues because they have more than one memory controller which means they have to register mce decoder many times. Cc: <stable@vger.kernel.org> # 3.2 and upper Signed-off-by: Chen Gong <gong.chen@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
2012-05-09 06:40:12 +07:00
mce_unregister_decode_chain(&i7_mce_dec);
}
module_init(i7core_init);
module_exit(i7core_exit);
MODULE_LICENSE("GPL");
MODULE_AUTHOR("Mauro Carvalho Chehab <mchehab@redhat.com>");
MODULE_AUTHOR("Red Hat Inc. (http://www.redhat.com)");
MODULE_DESCRIPTION("MC Driver for Intel i7 Core memory controllers - "
I7CORE_REVISION);
module_param(edac_op_state, int, 0444);
MODULE_PARM_DESC(edac_op_state, "EDAC Error Reporting state: 0=Poll,1=NMI");