summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2009-09-17 09:55:52 -0700
committerLinus Torvalds <torvalds@linux-foundation.org>2009-09-17 09:55:52 -0700
commitde55a8958f6e3ef5ce5f0971b80bd44bfcac7cf1 (patch)
tree6a29f442764bb13dbd6c2abe60e70f9f2cb1961a
parent79b520e87e1214cfa107bdc8528b5d6c055a8b82 (diff)
parent06724535f8fa26e78238bf8adfc9c81650a665f7 (diff)
downloadlinux-stable-de55a8958f6e3ef5ce5f0971b80bd44bfcac7cf1.tar.gz
linux-stable-de55a8958f6e3ef5ce5f0971b80bd44bfcac7cf1.tar.bz2
linux-stable-de55a8958f6e3ef5ce5f0971b80bd44bfcac7cf1.zip
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/bp/bp
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/bp/bp: amd64_edac: check NB MCE bank enable on the current node properly amd64_edac: Rewrite unganged mode code of f10_early_channel_count amd64_edac: cleanup amd64_check_ecc_enabled x86, EDAC: Provide function to return NodeId of a CPU amd64_edac: build driver only on AMD hardware
-rw-r--r--arch/x86/include/asm/processor.h2
-rw-r--r--arch/x86/kernel/cpu/amd.c10
-rw-r--r--drivers/edac/Kconfig2
-rw-r--r--drivers/edac/amd64_edac.c175
-rw-r--r--drivers/edac/edac_mce_amd.c2
5 files changed, 88 insertions, 103 deletions
diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h
index e08ea043e085..42a3f936dadc 100644
--- a/arch/x86/include/asm/processor.h
+++ b/arch/x86/include/asm/processor.h
@@ -1020,4 +1020,6 @@ extern void start_thread(struct pt_regs *regs, unsigned long new_ip,
extern int get_tsc_mode(unsigned long adr);
extern int set_tsc_mode(unsigned int val);
+extern int amd_get_nb_id(int cpu);
+
#endif /* _ASM_X86_PROCESSOR_H */
diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index 22a47c82f3c0..f32fa71ccf97 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -333,6 +333,16 @@ static void __cpuinit amd_detect_cmp(struct cpuinfo_x86 *c)
#endif
}
+int amd_get_nb_id(int cpu)
+{
+ int id = 0;
+#ifdef CONFIG_SMP
+ id = per_cpu(cpu_llc_id, cpu);
+#endif
+ return id;
+}
+EXPORT_SYMBOL_GPL(amd_get_nb_id);
+
static void __cpuinit srat_detect_node(struct cpuinfo_x86 *c)
{
#if defined(CONFIG_NUMA) && defined(CONFIG_X86_64)
diff --git a/drivers/edac/Kconfig b/drivers/edac/Kconfig
index 4339b1a879cd..a3ca18e2d7cf 100644
--- a/drivers/edac/Kconfig
+++ b/drivers/edac/Kconfig
@@ -59,7 +59,7 @@ config EDAC_MM_EDAC
config EDAC_AMD64
tristate "AMD64 (Opteron, Athlon64) K8, F10h, F11h"
- depends on EDAC_MM_EDAC && K8_NB && X86_64 && PCI
+ depends on EDAC_MM_EDAC && K8_NB && X86_64 && PCI && CPU_SUP_AMD
help
Support for error detection and correction on the AMD 64
Families of Memory Controllers (K8, F10h and F11h)
diff --git a/drivers/edac/amd64_edac.c b/drivers/edac/amd64_edac.c
index 173dc4a84166..4e551e63b6dc 100644
--- a/drivers/edac/amd64_edac.c
+++ b/drivers/edac/amd64_edac.c
@@ -1255,7 +1255,9 @@ static int k8_dbam_map_to_pages(struct amd64_pvt *pvt, int dram_map)
*/
static int f10_early_channel_count(struct amd64_pvt *pvt)
{
+ int dbams[] = { DBAM0, DBAM1 };
int err = 0, channels = 0;
+ int i, j;
u32 dbam;
err = pci_read_config_dword(pvt->dram_f2_ctl, F10_DCLR_0, &pvt->dclr0);
@@ -1288,46 +1290,19 @@ static int f10_early_channel_count(struct amd64_pvt *pvt)
* is more than just one DIMM present in unganged mode. Need to check
* both controllers since DIMMs can be placed in either one.
*/
- channels = 0;
- err = pci_read_config_dword(pvt->dram_f2_ctl, DBAM0, &dbam);
- if (err)
- goto err_reg;
-
- if (DBAM_DIMM(0, dbam) > 0)
- channels++;
- if (DBAM_DIMM(1, dbam) > 0)
- channels++;
- if (DBAM_DIMM(2, dbam) > 0)
- channels++;
- if (DBAM_DIMM(3, dbam) > 0)
- channels++;
-
- /* If more than 2 DIMMs are present, then we have 2 channels */
- if (channels > 2)
- channels = 2;
- else if (channels == 0) {
- /* No DIMMs on DCT0, so look at DCT1 */
- err = pci_read_config_dword(pvt->dram_f2_ctl, DBAM1, &dbam);
+ for (i = 0; i < ARRAY_SIZE(dbams); i++) {
+ err = pci_read_config_dword(pvt->dram_f2_ctl, dbams[i], &dbam);
if (err)
goto err_reg;
- if (DBAM_DIMM(0, dbam) > 0)
- channels++;
- if (DBAM_DIMM(1, dbam) > 0)
- channels++;
- if (DBAM_DIMM(2, dbam) > 0)
- channels++;
- if (DBAM_DIMM(3, dbam) > 0)
- channels++;
-
- if (channels > 2)
- channels = 2;
+ for (j = 0; j < 4; j++) {
+ if (DBAM_DIMM(j, dbam) > 0) {
+ channels++;
+ break;
+ }
+ }
}
- /* If we found ALL 0 values, then assume just ONE DIMM-ONE Channel */
- if (channels == 0)
- channels = 1;
-
debugf0("MCT channel count: %d\n", channels);
return channels;
@@ -2766,30 +2741,53 @@ static void amd64_restore_ecc_error_reporting(struct amd64_pvt *pvt)
wrmsr_on_cpus(cpumask, K8_MSR_MCGCTL, msrs);
}
-static void check_mcg_ctl(void *ret)
+/* get all cores on this DCT */
+static void get_cpus_on_this_dct_cpumask(cpumask_t *mask, int nid)
{
- u64 msr_val = 0;
- u8 nbe;
-
- rdmsrl(MSR_IA32_MCG_CTL, msr_val);
- nbe = msr_val & K8_MSR_MCGCTL_NBE;
-
- debugf0("core: %u, MCG_CTL: 0x%llx, NB MSR is %s\n",
- raw_smp_processor_id(), msr_val,
- (nbe ? "enabled" : "disabled"));
+ int cpu;
- if (!nbe)
- *(int *)ret = 0;
+ for_each_online_cpu(cpu)
+ if (amd_get_nb_id(cpu) == nid)
+ cpumask_set_cpu(cpu, mask);
}
/* check MCG_CTL on all the cpus on this node */
-static int amd64_mcg_ctl_enabled_on_cpus(const cpumask_t *mask)
+static bool amd64_nb_mce_bank_enabled_on_node(int nid)
{
- int ret = 1;
- preempt_disable();
- smp_call_function_many(mask, check_mcg_ctl, &ret, 1);
- preempt_enable();
+ cpumask_t mask;
+ struct msr *msrs;
+ int cpu, nbe, idx = 0;
+ bool ret = false;
+ cpumask_clear(&mask);
+
+ get_cpus_on_this_dct_cpumask(&mask, nid);
+
+ msrs = kzalloc(sizeof(struct msr) * cpumask_weight(&mask), GFP_KERNEL);
+ if (!msrs) {
+ amd64_printk(KERN_WARNING, "%s: error allocating msrs\n",
+ __func__);
+ return false;
+ }
+
+ rdmsr_on_cpus(&mask, MSR_IA32_MCG_CTL, msrs);
+
+ for_each_cpu(cpu, &mask) {
+ nbe = msrs[idx].l & K8_MSR_MCGCTL_NBE;
+
+ debugf0("core: %u, MCG_CTL: 0x%llx, NB MSR is %s\n",
+ cpu, msrs[idx].q,
+ (nbe ? "enabled" : "disabled"));
+
+ if (!nbe)
+ goto out;
+
+ idx++;
+ }
+ ret = true;
+
+out:
+ kfree(msrs);
return ret;
}
@@ -2799,71 +2797,46 @@ static int amd64_mcg_ctl_enabled_on_cpus(const cpumask_t *mask)
* the memory system completely. A command line option allows to force-enable
* hardware ECC later in amd64_enable_ecc_error_reporting().
*/
+static const char *ecc_warning =
+ "WARNING: ECC is disabled by BIOS. Module will NOT be loaded.\n"
+ " Either Enable ECC in the BIOS, or set 'ecc_enable_override'.\n"
+ " Also, use of the override can cause unknown side effects.\n";
+
static int amd64_check_ecc_enabled(struct amd64_pvt *pvt)
{
u32 value;
- int err = 0, ret = 0;
+ int err = 0;
u8 ecc_enabled = 0;
+ bool nb_mce_en = false;
err = pci_read_config_dword(pvt->misc_f3_ctl, K8_NBCFG, &value);
if (err)
debugf0("Reading K8_NBCTL failed\n");
ecc_enabled = !!(value & K8_NBCFG_ECC_ENABLE);
+ if (!ecc_enabled)
+ amd64_printk(KERN_WARNING, "This node reports that Memory ECC "
+ "is currently disabled, set F3x%x[22] (%s).\n",
+ K8_NBCFG, pci_name(pvt->misc_f3_ctl));
+ else
+ amd64_printk(KERN_INFO, "ECC is enabled by BIOS.\n");
- ret = amd64_mcg_ctl_enabled_on_cpus(cpumask_of_node(pvt->mc_node_id));
-
- debugf0("K8_NBCFG=0x%x, DRAM ECC is %s\n", value,
- (value & K8_NBCFG_ECC_ENABLE ? "enabled" : "disabled"));
-
- if (!ecc_enabled || !ret) {
- if (!ecc_enabled) {
- amd64_printk(KERN_WARNING, "This node reports that "
- "Memory ECC is currently "
- "disabled.\n");
+ nb_mce_en = amd64_nb_mce_bank_enabled_on_node(pvt->mc_node_id);
+ if (!nb_mce_en)
+ amd64_printk(KERN_WARNING, "NB MCE bank disabled, set MSR "
+ "0x%08x[4] on node %d to enable.\n",
+ MSR_IA32_MCG_CTL, pvt->mc_node_id);
- amd64_printk(KERN_WARNING, "bit 0x%lx in register "
- "F3x%x of the MISC_CONTROL device (%s) "
- "should be enabled\n", K8_NBCFG_ECC_ENABLE,
- K8_NBCFG, pci_name(pvt->misc_f3_ctl));
- }
- if (!ret) {
- amd64_printk(KERN_WARNING, "bit 0x%016lx in MSR 0x%08x "
- "of node %d should be enabled\n",
- K8_MSR_MCGCTL_NBE, MSR_IA32_MCG_CTL,
- pvt->mc_node_id);
- }
+ if (!ecc_enabled || !nb_mce_en) {
if (!ecc_enable_override) {
- amd64_printk(KERN_WARNING, "WARNING: ECC is NOT "
- "currently enabled by the BIOS. Module "
- "will NOT be loaded.\n"
- " Either Enable ECC in the BIOS, "
- "or use the 'ecc_enable_override' "
- "parameter.\n"
- " Might be a BIOS bug, if BIOS says "
- "ECC is enabled\n"
- " Use of the override can cause "
- "unknown side effects.\n");
- ret = -ENODEV;
- } else
- /*
- * enable further driver loading if ECC enable is
- * overridden.
- */
- ret = 0;
- } else {
- amd64_printk(KERN_INFO,
- "ECC is enabled by BIOS, Proceeding "
- "with EDAC module initialization\n");
-
- /* Signal good ECC status */
- ret = 0;
-
+ amd64_printk(KERN_WARNING, "%s", ecc_warning);
+ return -ENODEV;
+ }
+ } else
/* CLEAR the override, since BIOS controlled it */
ecc_enable_override = 0;
- }
- return ret;
+ return 0;
}
struct mcidev_sysfs_attribute sysfs_attrs[ARRAY_SIZE(amd64_dbg_attrs) +
diff --git a/drivers/edac/edac_mce_amd.c b/drivers/edac/edac_mce_amd.c
index c8ca7136dacc..0c21c370c9dd 100644
--- a/drivers/edac/edac_mce_amd.c
+++ b/drivers/edac/edac_mce_amd.c
@@ -405,7 +405,7 @@ void decode_mce(struct mce *m)
regs.nbsh = (u32)(m->status >> 32);
regs.nbeal = (u32) m->addr;
regs.nbeah = (u32)(m->addr >> 32);
- node = per_cpu(cpu_llc_id, m->extcpu);
+ node = amd_get_nb_id(m->extcpu);
amd_decode_nb_mce(node, &regs, 1);
break;