2013-04-30 08:38:14

by Ingo Molnar

[permalink] [raw]
Subject: [GIT PULL] x86/ras changes for v3.10

Linus,

Please pull the latest x86-ras-for-linus git tree from:

git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86-ras-for-linus

HEAD: 5379f8c0d72cab43bbe6d974ceb3ad84dddc2b8e Merge tag 'edac_amd_f16h' of git://git.kernel.org/pub/scm/linux/kernel/git/ras/ras into x86/ras

This merge:

- Adds an Intel CMCI hotplug fix

- Adds AMD family 16h EDAC support

- Makes the AMD MCE banks code more flexible for virtual environments

out-of-topic modifications in x86-ras-for-linus:
------------------------------------------------
drivers/edac/amd64_edac.c # 94c1acf: amd64_edac: Add Family 16h suppor
drivers/edac/amd64_edac.h # 94c1acf: amd64_edac: Add Family 16h suppor
include/linux/pci_ids.h # 94c1acf: amd64_edac: Add Family 16h suppor

Thanks,

Ingo

------------------>
Aravind Gopalakrishnan (1):
amd64_edac: Add Family 16h support

Boris Ostrovsky (2):
x86, MCE, AMD: Replace shared_bank array with is_shared_bank() helper
x86, MCE, AMD: Use MCG_CAP MSR to find out number of banks on AMD

Srivatsa S. Bhat (1):
x86/mce: Rework cmci_rediscover() to play well with CPU hotplug


arch/x86/include/asm/mce.h | 4 +--
arch/x86/kernel/amd_nb.c | 3 +-
arch/x86/kernel/cpu/mcheck/mce.c | 2 +-
arch/x86/kernel/cpu/mcheck/mce_amd.c | 39 ++++++++++++--------
arch/x86/kernel/cpu/mcheck/mce_intel.c | 25 +++----------
drivers/edac/amd64_edac.c | 65 +++++++++++++++++++++++++++++++++-
drivers/edac/amd64_edac.h | 4 ++-
include/linux/pci_ids.h | 2 ++
8 files changed, 103 insertions(+), 41 deletions(-)

diff --git a/arch/x86/include/asm/mce.h b/arch/x86/include/asm/mce.h
index f4076af..fa5f71e 100644
--- a/arch/x86/include/asm/mce.h
+++ b/arch/x86/include/asm/mce.h
@@ -146,13 +146,13 @@ DECLARE_PER_CPU(struct device *, mce_device);
void mce_intel_feature_init(struct cpuinfo_x86 *c);
void cmci_clear(void);
void cmci_reenable(void);
-void cmci_rediscover(int dying);
+void cmci_rediscover(void);
void cmci_recheck(void);
#else
static inline void mce_intel_feature_init(struct cpuinfo_x86 *c) { }
static inline void cmci_clear(void) {}
static inline void cmci_reenable(void) {}
-static inline void cmci_rediscover(int dying) {}
+static inline void cmci_rediscover(void) {}
static inline void cmci_recheck(void) {}
#endif

diff --git a/arch/x86/kernel/amd_nb.c b/arch/x86/kernel/amd_nb.c
index aadf335..f1d0a14 100644
--- a/arch/x86/kernel/amd_nb.c
+++ b/arch/x86/kernel/amd_nb.c
@@ -20,12 +20,14 @@ const struct pci_device_id amd_nb_misc_ids[] = {
{ PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_10H_NB_MISC) },
{ PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_15H_NB_F3) },
{ PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_15H_M10H_F3) },
+ { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_16H_NB_F3) },
{}
};
EXPORT_SYMBOL(amd_nb_misc_ids);

static struct pci_device_id amd_nb_link_ids[] = {
{ PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_15H_NB_F4) },
+ { PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_16H_NB_F4) },
{}
};

@@ -81,7 +83,6 @@ int amd_cache_northbridges(void)
next_northbridge(link, amd_nb_link_ids);
}

- /* some CPU families (e.g. family 0x11) do not support GART */
if (boot_cpu_data.x86 == 0xf || boot_cpu_data.x86 == 0x10 ||
boot_cpu_data.x86 == 0x15)
amd_northbridges.flags |= AMD_NB_GART;
diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c
index 7bc1263..9239504 100644
--- a/arch/x86/kernel/cpu/mcheck/mce.c
+++ b/arch/x86/kernel/cpu/mcheck/mce.c
@@ -2358,7 +2358,7 @@ mce_cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu)

if (action == CPU_POST_DEAD) {
/* intentionally ignoring frozen here */
- cmci_rediscover(cpu);
+ cmci_rediscover();
}

return NOTIFY_OK;
diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c
index 1ac581f..9cb5276 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c
@@ -33,7 +33,6 @@
#include <asm/mce.h>
#include <asm/msr.h>

-#define NR_BANKS 6
#define NR_BLOCKS 9
#define THRESHOLD_MAX 0xFFF
#define INT_TYPE_APIC 0x00020000
@@ -57,12 +56,7 @@ static const char * const th_names[] = {
"execution_unit",
};

-static DEFINE_PER_CPU(struct threshold_bank * [NR_BANKS], threshold_banks);
-
-static unsigned char shared_bank[NR_BANKS] = {
- 0, 0, 0, 0, 1
-};
-
+static DEFINE_PER_CPU(struct threshold_bank **, threshold_banks);
static DEFINE_PER_CPU(unsigned char, bank_map); /* see which banks are on */

static void amd_threshold_interrupt(void);
@@ -79,6 +73,12 @@ struct thresh_restart {
u16 old_limit;
};

+static inline bool is_shared_bank(int bank)
+{
+ /* Bank 4 is for northbridge reporting and is thus shared */
+ return (bank == 4);
+}
+
static const char * const bank4_names(struct threshold_block *b)
{
switch (b->address) {
@@ -214,7 +214,7 @@ void mce_amd_feature_init(struct cpuinfo_x86 *c)
unsigned int bank, block;
int offset = -1;

- for (bank = 0; bank < NR_BANKS; ++bank) {
+ for (bank = 0; bank < mca_cfg.banks; ++bank) {
for (block = 0; block < NR_BLOCKS; ++block) {
if (block == 0)
address = MSR_IA32_MC0_MISC + bank * 4;
@@ -276,7 +276,7 @@ static void amd_threshold_interrupt(void)
mce_setup(&m);

/* assume first bank caused it */
- for (bank = 0; bank < NR_BANKS; ++bank) {
+ for (bank = 0; bank < mca_cfg.banks; ++bank) {
if (!(per_cpu(bank_map, m.cpu) & (1 << bank)))
continue;
for (block = 0; block < NR_BLOCKS; ++block) {
@@ -467,7 +467,7 @@ static __cpuinit int allocate_threshold_blocks(unsigned int cpu,
u32 low, high;
int err;

- if ((bank >= NR_BANKS) || (block >= NR_BLOCKS))
+ if ((bank >= mca_cfg.banks) || (block >= NR_BLOCKS))
return 0;

if (rdmsr_safe_on_cpu(cpu, address, &low, &high))
@@ -575,7 +575,7 @@ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank)
const char *name = th_names[bank];
int err = 0;

- if (shared_bank[bank]) {
+ if (is_shared_bank(bank)) {
nb = node_to_amd_nb(amd_get_nb_id(cpu));

/* threshold descriptor already initialized on this node? */
@@ -609,7 +609,7 @@ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank)

per_cpu(threshold_banks, cpu)[bank] = b;

- if (shared_bank[bank]) {
+ if (is_shared_bank(bank)) {
atomic_set(&b->cpus, 1);

/* nb is already initialized, see above */
@@ -635,9 +635,17 @@ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank)
static __cpuinit int threshold_create_device(unsigned int cpu)
{
unsigned int bank;
+ struct threshold_bank **bp;
int err = 0;

- for (bank = 0; bank < NR_BANKS; ++bank) {
+ bp = kzalloc(sizeof(struct threshold_bank *) * mca_cfg.banks,
+ GFP_KERNEL);
+ if (!bp)
+ return -ENOMEM;
+
+ per_cpu(threshold_banks, cpu) = bp;
+
+ for (bank = 0; bank < mca_cfg.banks; ++bank) {
if (!(per_cpu(bank_map, cpu) & (1 << bank)))
continue;
err = threshold_create_bank(cpu, bank);
@@ -691,7 +699,7 @@ static void threshold_remove_bank(unsigned int cpu, int bank)
if (!b->blocks)
goto free_out;

- if (shared_bank[bank]) {
+ if (is_shared_bank(bank)) {
if (!atomic_dec_and_test(&b->cpus)) {
__threshold_remove_blocks(b);
per_cpu(threshold_banks, cpu)[bank] = NULL;
@@ -719,11 +727,12 @@ static void threshold_remove_device(unsigned int cpu)
{
unsigned int bank;

- for (bank = 0; bank < NR_BANKS; ++bank) {
+ for (bank = 0; bank < mca_cfg.banks; ++bank) {
if (!(per_cpu(bank_map, cpu) & (1 << bank)))
continue;
threshold_remove_bank(cpu, bank);
}
+ kfree(per_cpu(threshold_banks, cpu));
}

/* get notified when a cpu comes on/off */
diff --git a/arch/x86/kernel/cpu/mcheck/mce_intel.c b/arch/x86/kernel/cpu/mcheck/mce_intel.c
index 402c454..ae1697c 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_intel.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_intel.c
@@ -285,39 +285,24 @@ void cmci_clear(void)
raw_spin_unlock_irqrestore(&cmci_discover_lock, flags);
}

-static long cmci_rediscover_work_func(void *arg)
+static void cmci_rediscover_work_func(void *arg)
{
int banks;

/* Recheck banks in case CPUs don't all have the same */
if (cmci_supported(&banks))
cmci_discover(banks);
-
- return 0;
}

-/*
- * After a CPU went down cycle through all the others and rediscover
- * Must run in process context.
- */
-void cmci_rediscover(int dying)
+/* After a CPU went down cycle through all the others and rediscover */
+void cmci_rediscover(void)
{
- int cpu, banks;
+ int banks;

if (!cmci_supported(&banks))
return;

- for_each_online_cpu(cpu) {
- if (cpu == dying)
- continue;
-
- if (cpu == smp_processor_id()) {
- cmci_rediscover_work_func(NULL);
- continue;
- }
-
- work_on_cpu(cpu, cmci_rediscover_work_func, NULL);
- }
+ on_each_cpu(cmci_rediscover_work_func, NULL, 1);
}

/*
diff --git a/drivers/edac/amd64_edac.c b/drivers/edac/amd64_edac.c
index e1d13c4..8b6a034 100644
--- a/drivers/edac/amd64_edac.c
+++ b/drivers/edac/amd64_edac.c
@@ -98,6 +98,7 @@ int __amd64_write_pci_cfg_dword(struct pci_dev *pdev, int offset,
*
* F15h: we select which DCT we access using F1x10C[DctCfgSel]
*
+ * F16h: has only 1 DCT
*/
static int k8_read_dct_pci_cfg(struct amd64_pvt *pvt, int addr, u32 *val,
const char *func)
@@ -340,6 +341,27 @@ static void get_cs_base_and_mask(struct amd64_pvt *pvt, int csrow, u8 dct,
base_bits = GENMASK(21, 31) | GENMASK(9, 15);
mask_bits = GENMASK(21, 29) | GENMASK(9, 15);
addr_shift = 4;
+
+ /*
+ * F16h needs two addr_shift values: 8 for high and 6 for low
+ * (cf. F16h BKDG).
+ */
+ } else if (boot_cpu_data.x86 == 0x16) {
+ csbase = pvt->csels[dct].csbases[csrow];
+ csmask = pvt->csels[dct].csmasks[csrow >> 1];
+
+ *base = (csbase & GENMASK(5, 15)) << 6;
+ *base |= (csbase & GENMASK(19, 30)) << 8;
+
+ *mask = ~0ULL;
+ /* poke holes for the csmask */
+ *mask &= ~((GENMASK(5, 15) << 6) |
+ (GENMASK(19, 30) << 8));
+
+ *mask |= (csmask & GENMASK(5, 15)) << 6;
+ *mask |= (csmask & GENMASK(19, 30)) << 8;
+
+ return;
} else {
csbase = pvt->csels[dct].csbases[csrow];
csmask = pvt->csels[dct].csmasks[csrow >> 1];
@@ -1150,6 +1172,21 @@ static int f15_dbam_to_chip_select(struct amd64_pvt *pvt, u8 dct,
return ddr3_cs_size(cs_mode, false);
}

+/*
+ * F16h has only limited cs_modes
+ */
+static int f16_dbam_to_chip_select(struct amd64_pvt *pvt, u8 dct,
+ unsigned cs_mode)
+{
+ WARN_ON(cs_mode > 12);
+
+ if (cs_mode == 6 || cs_mode == 8 ||
+ cs_mode == 9 || cs_mode == 12)
+ return -1;
+ else
+ return ddr3_cs_size(cs_mode, false);
+}
+
static void read_dram_ctl_register(struct amd64_pvt *pvt)
{

@@ -1587,6 +1624,17 @@ static struct amd64_family_type amd64_family_types[] = {
.read_dct_pci_cfg = f15_read_dct_pci_cfg,
}
},
+ [F16_CPUS] = {
+ .ctl_name = "F16h",
+ .f1_id = PCI_DEVICE_ID_AMD_16H_NB_F1,
+ .f3_id = PCI_DEVICE_ID_AMD_16H_NB_F3,
+ .ops = {
+ .early_channel_count = f1x_early_channel_count,
+ .map_sysaddr_to_csrow = f1x_map_sysaddr_to_csrow,
+ .dbam_to_cs = f16_dbam_to_chip_select,
+ .read_dct_pci_cfg = f10_read_dct_pci_cfg,
+ }
+ },
};

/*
@@ -1939,7 +1987,9 @@ static void read_mc_regs(struct amd64_pvt *pvt)

if (c->x86 >= 0x10) {
amd64_read_pci_cfg(pvt->F3, EXT_NB_MCA_CFG, &tmp);
- amd64_read_dct_pci_cfg(pvt, DBAM1, &pvt->dbam1);
+ if (c->x86 != 0x16)
+ /* F16h has only DCT0 */
+ amd64_read_dct_pci_cfg(pvt, DBAM1, &pvt->dbam1);

/* F10h, revD and later can do x8 ECC too */
if ((c->x86 > 0x10 || c->x86_model > 7) && tmp & BIT(25))
@@ -2356,6 +2406,11 @@ static struct amd64_family_type *amd64_per_family_init(struct amd64_pvt *pvt)
pvt->ops = &amd64_family_types[F15_CPUS].ops;
break;

+ case 0x16:
+ fam_type = &amd64_family_types[F16_CPUS];
+ pvt->ops = &amd64_family_types[F16_CPUS].ops;
+ break;
+
default:
amd64_err("Unsupported family!\n");
return NULL;
@@ -2581,6 +2636,14 @@ static DEFINE_PCI_DEVICE_TABLE(amd64_pci_table) = {
.class = 0,
.class_mask = 0,
},
+ {
+ .vendor = PCI_VENDOR_ID_AMD,
+ .device = PCI_DEVICE_ID_AMD_16H_NB_F2,
+ .subvendor = PCI_ANY_ID,
+ .subdevice = PCI_ANY_ID,
+ .class = 0,
+ .class_mask = 0,
+ },

{0, }
};
diff --git a/drivers/edac/amd64_edac.h b/drivers/edac/amd64_edac.h
index 35637d8..2c6f113 100644
--- a/drivers/edac/amd64_edac.h
+++ b/drivers/edac/amd64_edac.h
@@ -172,7 +172,8 @@
*/
#define PCI_DEVICE_ID_AMD_15H_NB_F1 0x1601
#define PCI_DEVICE_ID_AMD_15H_NB_F2 0x1602
-
+#define PCI_DEVICE_ID_AMD_16H_NB_F1 0x1531
+#define PCI_DEVICE_ID_AMD_16H_NB_F2 0x1532

/*
* Function 1 - Address Map
@@ -296,6 +297,7 @@ enum amd_families {
K8_CPUS = 0,
F10_CPUS,
F15_CPUS,
+ F16_CPUS,
NUM_FAMILIES,
};

diff --git a/include/linux/pci_ids.h b/include/linux/pci_ids.h
index f11c1c2..9b3b858 100644
--- a/include/linux/pci_ids.h
+++ b/include/linux/pci_ids.h
@@ -524,6 +524,8 @@
#define PCI_DEVICE_ID_AMD_15H_NB_F3 0x1603
#define PCI_DEVICE_ID_AMD_15H_NB_F4 0x1604
#define PCI_DEVICE_ID_AMD_15H_NB_F5 0x1605
+#define PCI_DEVICE_ID_AMD_16H_NB_F3 0x1533
+#define PCI_DEVICE_ID_AMD_16H_NB_F4 0x1534
#define PCI_DEVICE_ID_AMD_CNB17H_F3 0x1703
#define PCI_DEVICE_ID_AMD_LANCE 0x2000
#define PCI_DEVICE_ID_AMD_LANCE_HOME 0x2001