Commit 0f959e19 authored by Borislav Petkov's avatar Borislav Petkov

Merge branch 'edac-ghes' into edac-for-next

parents 17ed808a e370f886
...@@ -15,9 +15,7 @@ ...@@ -15,9 +15,7 @@
#include "edac_module.h" #include "edac_module.h"
#include <ras/ras_event.h> #include <ras/ras_event.h>
struct ghes_edac_pvt { struct ghes_pvt {
struct list_head list;
struct ghes *ghes;
struct mem_ctl_info *mci; struct mem_ctl_info *mci;
/* Buffers for the error handling routine */ /* Buffers for the error handling routine */
...@@ -32,7 +30,16 @@ static refcount_t ghes_refcount = REFCOUNT_INIT(0); ...@@ -32,7 +30,16 @@ static refcount_t ghes_refcount = REFCOUNT_INIT(0);
* also provides the necessary (implicit) memory barrier for the SMP * also provides the necessary (implicit) memory barrier for the SMP
* case to make the pointer visible on another CPU. * case to make the pointer visible on another CPU.
*/ */
static struct ghes_edac_pvt *ghes_pvt; static struct ghes_pvt *ghes_pvt;
/*
* This driver's representation of the system hardware, as collected
* from DMI.
*/
struct ghes_hw_desc {
int num_dimms;
struct dimm_info *dimms;
} ghes_hw;
/* GHES registration mutex */ /* GHES registration mutex */
static DEFINE_MUTEX(ghes_reg_mutex); static DEFINE_MUTEX(ghes_reg_mutex);
...@@ -74,44 +81,35 @@ struct memdev_dmi_entry { ...@@ -74,44 +81,35 @@ struct memdev_dmi_entry {
u16 conf_mem_clk_speed; u16 conf_mem_clk_speed;
} __attribute__((__packed__)); } __attribute__((__packed__));
struct ghes_edac_dimm_fill { static struct dimm_info *find_dimm_by_handle(struct mem_ctl_info *mci, u16 handle)
struct mem_ctl_info *mci;
unsigned int count;
};
static void ghes_edac_count_dimms(const struct dmi_header *dh, void *arg)
{
int *num_dimm = arg;
if (dh->type == DMI_ENTRY_MEM_DEVICE)
(*num_dimm)++;
}
static int get_dimm_smbios_index(struct mem_ctl_info *mci, u16 handle)
{ {
struct dimm_info *dimm; struct dimm_info *dimm;
mci_for_each_dimm(mci, dimm) { mci_for_each_dimm(mci, dimm) {
if (dimm->smbios_handle == handle) if (dimm->smbios_handle == handle)
return dimm->idx; return dimm;
} }
return -1; return NULL;
} }
static void ghes_edac_dmidecode(const struct dmi_header *dh, void *arg) static void dimm_setup_label(struct dimm_info *dimm, u16 handle)
{ {
struct ghes_edac_dimm_fill *dimm_fill = arg; const char *bank = NULL, *device = NULL;
struct mem_ctl_info *mci = dimm_fill->mci;
if (dh->type == DMI_ENTRY_MEM_DEVICE) { dmi_memdev_name(handle, &bank, &device);
struct memdev_dmi_entry *entry = (struct memdev_dmi_entry *)dh;
struct dimm_info *dimm = edac_get_dimm(mci, dimm_fill->count, 0, 0); /* both strings must be non-zero */
if (bank && *bank && device && *device)
snprintf(dimm->label, sizeof(dimm->label), "%s %s", bank, device);
}
static void assign_dmi_dimm_info(struct dimm_info *dimm, struct memdev_dmi_entry *entry)
{
u16 rdr_mask = BIT(7) | BIT(13); u16 rdr_mask = BIT(7) | BIT(13);
if (entry->size == 0xffff) { if (entry->size == 0xffff) {
pr_info("Can't get DIMM%i size\n", pr_info("Can't get DIMM%i size\n", dimm->idx);
dimm_fill->count);
dimm->nr_pages = MiB_TO_PAGES(32);/* Unknown */ dimm->nr_pages = MiB_TO_PAGES(32);/* Unknown */
} else if (entry->size == 0x7fff) { } else if (entry->size == 0x7fff) {
dimm->nr_pages = MiB_TO_PAGES(entry->extended_size); dimm->nr_pages = MiB_TO_PAGES(entry->extended_size);
...@@ -179,13 +177,11 @@ static void ghes_edac_dmidecode(const struct dmi_header *dh, void *arg) ...@@ -179,13 +177,11 @@ static void ghes_edac_dmidecode(const struct dmi_header *dh, void *arg)
dimm->dtype = DEV_UNKNOWN; dimm->dtype = DEV_UNKNOWN;
dimm->grain = 128; /* Likely, worse case */ dimm->grain = 128; /* Likely, worse case */
/* dimm_setup_label(dimm, entry->handle);
* FIXME: It shouldn't be hard to also fill the DIMM labels
*/
if (dimm->nr_pages) { if (dimm->nr_pages) {
edac_dbg(1, "DIMM%i: %s size = %d MB%s\n", edac_dbg(1, "DIMM%i: %s size = %d MB%s\n",
dimm_fill->count, edac_mem_types[dimm->mtype], dimm->idx, edac_mem_types[dimm->mtype],
PAGES_TO_MiB(dimm->nr_pages), PAGES_TO_MiB(dimm->nr_pages),
(dimm->edac_mode != EDAC_NONE) ? "(ECC)" : ""); (dimm->edac_mode != EDAC_NONE) ? "(ECC)" : "");
edac_dbg(2, "\ttype %d, detail 0x%02x, width %d(total %d)\n", edac_dbg(2, "\ttype %d, detail 0x%02x, width %d(total %d)\n",
...@@ -194,16 +190,56 @@ static void ghes_edac_dmidecode(const struct dmi_header *dh, void *arg) ...@@ -194,16 +190,56 @@ static void ghes_edac_dmidecode(const struct dmi_header *dh, void *arg)
} }
dimm->smbios_handle = entry->handle; dimm->smbios_handle = entry->handle;
}
dimm_fill->count++; static void enumerate_dimms(const struct dmi_header *dh, void *arg)
{
struct memdev_dmi_entry *entry = (struct memdev_dmi_entry *)dh;
struct ghes_hw_desc *hw = (struct ghes_hw_desc *)arg;
struct dimm_info *d;
if (dh->type != DMI_ENTRY_MEM_DEVICE)
return;
/* Enlarge the array with additional 16 */
if (!hw->num_dimms || !(hw->num_dimms % 16)) {
struct dimm_info *new;
new = krealloc(hw->dimms, (hw->num_dimms + 16) * sizeof(struct dimm_info),
GFP_KERNEL);
if (!new) {
WARN_ON_ONCE(1);
return;
} }
hw->dimms = new;
}
d = &hw->dimms[hw->num_dimms];
d->idx = hw->num_dimms;
assign_dmi_dimm_info(d, entry);
hw->num_dimms++;
}
static void ghes_scan_system(void)
{
static bool scanned;
if (scanned)
return;
dmi_walk(enumerate_dimms, &ghes_hw);
scanned = true;
} }
void ghes_edac_report_mem_error(int sev, struct cper_sec_mem_err *mem_err) void ghes_edac_report_mem_error(int sev, struct cper_sec_mem_err *mem_err)
{ {
struct edac_raw_error_desc *e; struct edac_raw_error_desc *e;
struct mem_ctl_info *mci; struct mem_ctl_info *mci;
struct ghes_edac_pvt *pvt; struct ghes_pvt *pvt;
unsigned long flags; unsigned long flags;
char *p; char *p;
...@@ -228,7 +264,6 @@ void ghes_edac_report_mem_error(int sev, struct cper_sec_mem_err *mem_err) ...@@ -228,7 +264,6 @@ void ghes_edac_report_mem_error(int sev, struct cper_sec_mem_err *mem_err)
memset(e, 0, sizeof (*e)); memset(e, 0, sizeof (*e));
e->error_count = 1; e->error_count = 1;
e->grain = 1; e->grain = 1;
strcpy(e->label, "unknown label");
e->msg = pvt->msg; e->msg = pvt->msg;
e->other_detail = pvt->other_detail; e->other_detail = pvt->other_detail;
e->top_layer = -1; e->top_layer = -1;
...@@ -345,7 +380,7 @@ void ghes_edac_report_mem_error(int sev, struct cper_sec_mem_err *mem_err) ...@@ -345,7 +380,7 @@ void ghes_edac_report_mem_error(int sev, struct cper_sec_mem_err *mem_err)
p += sprintf(p, "bit_pos:%d ", mem_err->bit_pos); p += sprintf(p, "bit_pos:%d ", mem_err->bit_pos);
if (mem_err->validation_bits & CPER_MEM_VALID_MODULE_HANDLE) { if (mem_err->validation_bits & CPER_MEM_VALID_MODULE_HANDLE) {
const char *bank = NULL, *device = NULL; const char *bank = NULL, *device = NULL;
int index = -1; struct dimm_info *dimm;
dmi_memdev_name(mem_err->mem_dev_handle, &bank, &device); dmi_memdev_name(mem_err->mem_dev_handle, &bank, &device);
if (bank != NULL && device != NULL) if (bank != NULL && device != NULL)
...@@ -354,13 +389,18 @@ void ghes_edac_report_mem_error(int sev, struct cper_sec_mem_err *mem_err) ...@@ -354,13 +389,18 @@ void ghes_edac_report_mem_error(int sev, struct cper_sec_mem_err *mem_err)
p += sprintf(p, "DIMM DMI handle: 0x%.4x ", p += sprintf(p, "DIMM DMI handle: 0x%.4x ",
mem_err->mem_dev_handle); mem_err->mem_dev_handle);
index = get_dimm_smbios_index(mci, mem_err->mem_dev_handle); dimm = find_dimm_by_handle(mci, mem_err->mem_dev_handle);
if (index >= 0) if (dimm) {
e->top_layer = index; e->top_layer = dimm->idx;
strcpy(e->label, dimm->label);
}
} }
if (p > e->location) if (p > e->location)
*(p - 1) = '\0'; *(p - 1) = '\0';
if (!*e->label)
strcpy(e->label, "unknown memory");
/* All other fields are mapped on e->other_detail */ /* All other fields are mapped on e->other_detail */
p = pvt->other_detail; p = pvt->other_detail;
p += snprintf(p, sizeof(pvt->other_detail), p += snprintf(p, sizeof(pvt->other_detail),
...@@ -455,13 +495,12 @@ static struct acpi_platform_list plat_list[] = { ...@@ -455,13 +495,12 @@ static struct acpi_platform_list plat_list[] = {
int ghes_edac_register(struct ghes *ghes, struct device *dev) int ghes_edac_register(struct ghes *ghes, struct device *dev)
{ {
bool fake = false; bool fake = false;
int rc = 0, num_dimm = 0;
struct mem_ctl_info *mci; struct mem_ctl_info *mci;
struct ghes_edac_pvt *pvt; struct ghes_pvt *pvt;
struct edac_mc_layer layers[1]; struct edac_mc_layer layers[1];
struct ghes_edac_dimm_fill dimm_fill;
unsigned long flags; unsigned long flags;
int idx = -1; int idx = -1;
int rc = 0;
if (IS_ENABLED(CONFIG_X86)) { if (IS_ENABLED(CONFIG_X86)) {
/* Check if safe to enable on this system */ /* Check if safe to enable on this system */
...@@ -481,20 +520,19 @@ int ghes_edac_register(struct ghes *ghes, struct device *dev) ...@@ -481,20 +520,19 @@ int ghes_edac_register(struct ghes *ghes, struct device *dev)
if (refcount_inc_not_zero(&ghes_refcount)) if (refcount_inc_not_zero(&ghes_refcount))
goto unlock; goto unlock;
/* Get the number of DIMMs */ ghes_scan_system();
dmi_walk(ghes_edac_count_dimms, &num_dimm);
/* Check if we've got a bogus BIOS */ /* Check if we've got a bogus BIOS */
if (num_dimm == 0) { if (!ghes_hw.num_dimms) {
fake = true; fake = true;
num_dimm = 1; ghes_hw.num_dimms = 1;
} }
layers[0].type = EDAC_MC_LAYER_ALL_MEM; layers[0].type = EDAC_MC_LAYER_ALL_MEM;
layers[0].size = num_dimm; layers[0].size = ghes_hw.num_dimms;
layers[0].is_virt_csrow = true; layers[0].is_virt_csrow = true;
mci = edac_mc_alloc(0, ARRAY_SIZE(layers), layers, sizeof(struct ghes_edac_pvt)); mci = edac_mc_alloc(0, ARRAY_SIZE(layers), layers, sizeof(struct ghes_pvt));
if (!mci) { if (!mci) {
pr_info("Can't allocate memory for EDAC data\n"); pr_info("Can't allocate memory for EDAC data\n");
rc = -ENOMEM; rc = -ENOMEM;
...@@ -502,7 +540,6 @@ int ghes_edac_register(struct ghes *ghes, struct device *dev) ...@@ -502,7 +540,6 @@ int ghes_edac_register(struct ghes *ghes, struct device *dev)
} }
pvt = mci->pvt_info; pvt = mci->pvt_info;
pvt->ghes = ghes;
pvt->mci = mci; pvt->mci = mci;
mci->pdev = dev; mci->pdev = dev;
...@@ -523,13 +560,34 @@ int ghes_edac_register(struct ghes *ghes, struct device *dev) ...@@ -523,13 +560,34 @@ int ghes_edac_register(struct ghes *ghes, struct device *dev)
pr_info("So, the end result of using this driver varies from vendor to vendor.\n"); pr_info("So, the end result of using this driver varies from vendor to vendor.\n");
pr_info("If you find incorrect reports, please contact your hardware vendor\n"); pr_info("If you find incorrect reports, please contact your hardware vendor\n");
pr_info("to correct its BIOS.\n"); pr_info("to correct its BIOS.\n");
pr_info("This system has %d DIMM sockets.\n", num_dimm); pr_info("This system has %d DIMM sockets.\n", ghes_hw.num_dimms);
} }
if (!fake) { if (!fake) {
dimm_fill.count = 0; struct dimm_info *src, *dst;
dimm_fill.mci = mci; int i = 0;
dmi_walk(ghes_edac_dmidecode, &dimm_fill);
mci_for_each_dimm(mci, dst) {
src = &ghes_hw.dimms[i];
dst->idx = src->idx;
dst->smbios_handle = src->smbios_handle;
dst->nr_pages = src->nr_pages;
dst->mtype = src->mtype;
dst->edac_mode = src->edac_mode;
dst->dtype = src->dtype;
dst->grain = src->grain;
/*
* If no src->label, preserve default label assigned
* from EDAC core.
*/
if (strlen(src->label))
memcpy(dst->label, src->label, sizeof(src->label));
i++;
}
} else { } else {
struct dimm_info *dimm = edac_get_dimm(mci, 0, 0, 0); struct dimm_info *dimm = edac_get_dimm(mci, 0, 0, 0);
...@@ -542,7 +600,7 @@ int ghes_edac_register(struct ghes *ghes, struct device *dev) ...@@ -542,7 +600,7 @@ int ghes_edac_register(struct ghes *ghes, struct device *dev)
rc = edac_mc_add_mc(mci); rc = edac_mc_add_mc(mci);
if (rc < 0) { if (rc < 0) {
pr_info("Can't register at EDAC core\n"); pr_info("Can't register with the EDAC core\n");
edac_mc_free(mci); edac_mc_free(mci);
rc = -ENODEV; rc = -ENODEV;
goto unlock; goto unlock;
...@@ -556,6 +614,11 @@ int ghes_edac_register(struct ghes *ghes, struct device *dev) ...@@ -556,6 +614,11 @@ int ghes_edac_register(struct ghes *ghes, struct device *dev)
refcount_set(&ghes_refcount, 1); refcount_set(&ghes_refcount, 1);
unlock: unlock:
/* Not needed anymore */
kfree(ghes_hw.dimms);
ghes_hw.dimms = NULL;
mutex_unlock(&ghes_reg_mutex); mutex_unlock(&ghes_reg_mutex);
return rc; return rc;
......
...@@ -594,27 +594,6 @@ struct mem_ctl_info { ...@@ -594,27 +594,6 @@ struct mem_ctl_info {
? (mci)->dimms[(dimm)->idx + 1] \ ? (mci)->dimms[(dimm)->idx + 1] \
: NULL) : NULL)
/**
* edac_get_dimm_by_index - Get DIMM info at @index from a memory
* controller
*
* @mci: MC descriptor struct mem_ctl_info
* @index: index in the memory controller's DIMM array
*
* Returns a struct dimm_info * or NULL on failure.
*/
static inline struct dimm_info *
edac_get_dimm_by_index(struct mem_ctl_info *mci, int index)
{
if (index < 0 || index >= mci->tot_dimms)
return NULL;
if (WARN_ON_ONCE(mci->dimms[index]->idx != index))
return NULL;
return mci->dimms[index];
}
/** /**
* edac_get_dimm - Get DIMM info from a memory controller given by * edac_get_dimm - Get DIMM info from a memory controller given by
* [layer0,layer1,layer2] position * [layer0,layer1,layer2] position
...@@ -650,6 +629,12 @@ static inline struct dimm_info *edac_get_dimm(struct mem_ctl_info *mci, ...@@ -650,6 +629,12 @@ static inline struct dimm_info *edac_get_dimm(struct mem_ctl_info *mci,
if (mci->n_layers > 2) if (mci->n_layers > 2)
index = index * mci->layers[2].size + layer2; index = index * mci->layers[2].size + layer2;
return edac_get_dimm_by_index(mci, index); if (index < 0 || index >= mci->tot_dimms)
return NULL;
if (WARN_ON_ONCE(mci->dimms[index]->idx != index))
return NULL;
return mci->dimms[index];
} }
#endif /* _LINUX_EDAC_H_ */ #endif /* _LINUX_EDAC_H_ */
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment