[PATCH 0/9] x86, AMD, MCE: Thresholding cleanups

[PATCH 0/9] x86, AMD, MCE: Thresholding cleanups

[Borislav Petkov]

From: Borislav Petkov <borislav.petkov@amd.com>

Ok,

so those are a bunch of patches which give the MCE thresholding code a
hard look and a scrubbing to remove a couple of annoyances like sysfs
warnings when running CPU off-/online tests and the threshold_bank4 node
under /sys/devices/system/machinecheck/ is a symlink.

It also gives proper names to the thresholding banks instead of simply
enumerating them, see patch 8/9 for details.

It is tested on all our families >= K8 so if there are no objections,
I'd like to send it upwards next week and maybe even hit the upcoming
merge window, we'll see.

Thanks.

[PATCH 1/9] x86, amd_nb: Export model 0x10 and later PCI id

[Borislav Petkov]

From: Borislav Petkov <borislav.petkov@amd.com>

Add the F3 PCI id of F15h, model 0x10 to pci_ids.h and to the amd_nb
code which generates the list of northbridges on an AMD box. Shorten
define name while at it so that it fits into pci_ids.h.

Cc: Clemens Ladisch <clemens@ladisch.de>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Andreas Herrmann <andreas.herrmann3@amd.com>
Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
---
 arch/x86/kernel/amd_nb.c |    1 +
 drivers/hwmon/k10temp.c  |    5 +----
 include/linux/pci_ids.h  |    1 +
 3 files changed, 3 insertions(+), 4 deletions(-)

diff --git a/arch/x86/kernel/amd_nb.c b/arch/x86/kernel/amd_nb.c
index be16854591cc..153a0ee88fb1 100644
--- a/arch/x86/kernel/amd_nb.c
+++ b/arch/x86/kernel/amd_nb.c
@@ -16,6 +16,7 @@ const struct pci_device_id amd_nb_misc_ids[] = {
 	{ PCI_DEVICE(PCI_VENDOR_ID_AMD, PCI_DEVICE_ID_AMD_K8_NB_MISC) },
 	{ 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) },
 	{}
 };
 EXPORT_SYMBOL(amd_nb_misc_ids);
diff --git a/drivers/hwmon/k10temp.c b/drivers/hwmon/k10temp.c
index 307bb325dde9..723675ad8e4d 100644
--- a/drivers/hwmon/k10temp.c
+++ b/drivers/hwmon/k10temp.c
@@ -33,9 +33,6 @@ static bool force;
 module_param(force, bool, 0444);
 MODULE_PARM_DESC(force, "force loading on processors with erratum 319");
 
-/* PCI-IDs for Northbridge devices not used anywhere else */
-#define PCI_DEVICE_ID_AMD_15H_M10H_NB_F3	0x1403
-
 /* CPUID function 0x80000001, ebx */
 #define CPUID_PKGTYPE_MASK	0xf0000000
 #define CPUID_PKGTYPE_F		0x00000000
@@ -213,7 +210,7 @@ static DEFINE_PCI_DEVICE_TABLE(k10temp_id_table) = {
 	{ PCI_VDEVICE(AMD, PCI_DEVICE_ID_AMD_11H_NB_MISC) },
 	{ PCI_VDEVICE(AMD, PCI_DEVICE_ID_AMD_CNB17H_F3) },
 	{ PCI_VDEVICE(AMD, PCI_DEVICE_ID_AMD_15H_NB_F3) },
-	{ PCI_VDEVICE(AMD, PCI_DEVICE_ID_AMD_15H_M10H_NB_F3) },
+	{ PCI_VDEVICE(AMD, PCI_DEVICE_ID_AMD_15H_M10H_F3) },
 	{}
 };
 MODULE_DEVICE_TABLE(pci, k10temp_id_table);
diff --git a/include/linux/pci_ids.h b/include/linux/pci_ids.h
index 3329965ed63f..ba943aa2247e 100644
--- a/include/linux/pci_ids.h
+++ b/include/linux/pci_ids.h
@@ -517,6 +517,7 @@
 #define PCI_DEVICE_ID_AMD_11H_NB_DRAM	0x1302
 #define PCI_DEVICE_ID_AMD_11H_NB_MISC	0x1303
 #define PCI_DEVICE_ID_AMD_11H_NB_LINK	0x1304
+#define PCI_DEVICE_ID_AMD_15H_M10H_F3	0x1403
 #define PCI_DEVICE_ID_AMD_15H_NB_F0	0x1600
 #define PCI_DEVICE_ID_AMD_15H_NB_F1	0x1601
 #define PCI_DEVICE_ID_AMD_15H_NB_F2	0x1602
-- 
1.7.9.3.362.g71319

Re: [PATCH 1/9] x86, amd_nb: Export model 0x10 and later PCI id

[Clemens Ladisch]

Borislav Petkov wrote:
> Add the F3 PCI id of F15h, model 0x10 to pci_ids.h and to the amd_nb
> code which generates the list of northbridges on an AMD box. Shorten
> define name while at it so that it fits into pci_ids.h.
>
> Cc: Clemens Ladisch <clemens@ladisch.de>

Acked-by: Clemens Ladisch <clemens@ladisch.de>

> Cc: Bjorn Helgaas <bhelgaas@google.com>
> Acked-by: Andreas Herrmann <andreas.herrmann3@amd.com>
> Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
> ---
>  arch/x86/kernel/amd_nb.c |    1 +
>  drivers/hwmon/k10temp.c  |    5 +----
>  include/linux/pci_ids.h  |    1 +
>  3 files changed, 3 insertions(+), 4 deletions(-)

[PATCH 2/9] x86, MCE, AMD: Remove shared banks sysfs linking

[Borislav Petkov]

From: Borislav Petkov <borislav.petkov@amd.com>

The code used to create a symlink on all non-BSP cores of a node when
the MCi_MISCj bank is present once per node. (This is generally the
case with bank 4 on AMD). However, these sysfs links cause a bunch
of problems with cpu off-/onlining testing and are, as such, a bit
overengineered. IOW, there's nothing wrong with having normal sysfs
files for the shared banks since the corresponding MSRs are replicated
across each core anyway.

Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
---
 arch/x86/kernel/cpu/mcheck/mce_amd.c |  105 +++-------------------------------
 1 file changed, 7 insertions(+), 98 deletions(-)

diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c
index f4873a64f46d..2a5dd30f9969 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c
@@ -60,7 +60,6 @@ struct threshold_block {
 struct threshold_bank {
 	struct kobject		*kobj;
 	struct threshold_block	*blocks;
-	cpumask_var_t		cpus;
 };
 static DEFINE_PER_CPU(struct threshold_bank * [NR_BANKS], threshold_banks);
 
@@ -224,8 +223,6 @@ void mce_amd_feature_init(struct cpuinfo_x86 *c)
 
 			if (!block)
 				per_cpu(bank_map, cpu) |= (1 << bank);
-			if (shared_bank[bank] && c->cpu_core_id)
-				break;
 
 			memset(&b, 0, sizeof(b));
 			b.cpu			= cpu;
@@ -555,89 +552,35 @@ local_allocate_threshold_blocks(int cpu, unsigned int bank)
 					 MSR_IA32_MC0_MISC + bank * 4);
 }
 
-/* symlinks sibling shared banks to first core.  first core owns dir/files. */
 static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank)
 {
-	int i, err = 0;
-	struct threshold_bank *b = NULL;
 	struct device *dev = per_cpu(mce_device, cpu);
+	struct threshold_bank *b = NULL;
 	char name[32];
+	int err = 0;
 
 	sprintf(name, "threshold_bank%i", bank);
 
-#ifdef CONFIG_SMP
-	if (cpu_data(cpu).cpu_core_id && shared_bank[bank]) {	/* symlink */
-		i = cpumask_first(cpu_llc_shared_mask(cpu));
-
-		/* first core not up yet */
-		if (cpu_data(i).cpu_core_id)
-			goto out;
-
-		/* already linked */
-		if (per_cpu(threshold_banks, cpu)[bank])
-			goto out;
-
-		b = per_cpu(threshold_banks, i)[bank];
-
-		if (!b)
-			goto out;
-
-		err = sysfs_create_link(&dev->kobj, b->kobj, name);
-		if (err)
-			goto out;
-
-		cpumask_copy(b->cpus, cpu_llc_shared_mask(cpu));
-		per_cpu(threshold_banks, cpu)[bank] = b;
-
-		goto out;
-	}
-#endif
-
 	b = kzalloc(sizeof(struct threshold_bank), GFP_KERNEL);
 	if (!b) {
 		err = -ENOMEM;
 		goto out;
 	}
-	if (!zalloc_cpumask_var(&b->cpus, GFP_KERNEL)) {
-		kfree(b);
-		err = -ENOMEM;
-		goto out;
-	}
 
 	b->kobj = kobject_create_and_add(name, &dev->kobj);
-	if (!b->kobj)
+	if (!b->kobj) {
+		err = -EINVAL;
 		goto out_free;
-
-#ifndef CONFIG_SMP
-	cpumask_setall(b->cpus);
-#else
-	cpumask_set_cpu(cpu, b->cpus);
-#endif
+	}
 
 	per_cpu(threshold_banks, cpu)[bank] = b;
 
 	err = local_allocate_threshold_blocks(cpu, bank);
-	if (err)
-		goto out_free;
-
-	for_each_cpu(i, b->cpus) {
-		if (i == cpu)
-			continue;
-
-		dev = per_cpu(mce_device, i);
-		if (dev)
-			err = sysfs_create_link(&dev->kobj,b->kobj, name);
-		if (err)
-			goto out;
-
-		per_cpu(threshold_banks, i)[bank] = b;
-	}
-
-	goto out;
+	if (!err)
+		goto out;
 
 out_free:
 	per_cpu(threshold_banks, cpu)[bank] = NULL;
-	free_cpumask_var(b->cpus);
 	kfree(b);
 out:
 	return err;
@@ -660,12 +603,6 @@ static __cpuinit int threshold_create_device(unsigned int cpu)
 	return err;
 }
 
-/*
- * let's be hotplug friendly.
- * in case of multiple core processors, the first core always takes ownership
- *   of shared sysfs dir/files, and rest of the cores will be symlinked to it.
- */
-
 static void deallocate_threshold_block(unsigned int cpu,
 						 unsigned int bank)
 {
@@ -689,9 +626,6 @@ static void deallocate_threshold_block(unsigned int cpu,
 static void threshold_remove_bank(unsigned int cpu, int bank)
 {
 	struct threshold_bank *b;
-	struct device *dev;
-	char name[32];
-	int i = 0;
 
 	b = per_cpu(threshold_banks, cpu)[bank];
 	if (!b)
@@ -699,36 +633,11 @@ static void threshold_remove_bank(unsigned int cpu, int bank)
 	if (!b->blocks)
 		goto free_out;
 
-	sprintf(name, "threshold_bank%i", bank);
-
-#ifdef CONFIG_SMP
-	/* sibling symlink */
-	if (shared_bank[bank] && b->blocks->cpu != cpu) {
-		dev = per_cpu(mce_device, cpu);
-		sysfs_remove_link(&dev->kobj, name);
-		per_cpu(threshold_banks, cpu)[bank] = NULL;
-
-		return;
-	}
-#endif
-
-	/* remove all sibling symlinks before unregistering */
-	for_each_cpu(i, b->cpus) {
-		if (i == cpu)
-			continue;
-
-		dev = per_cpu(mce_device, i);
-		if (dev)
-			sysfs_remove_link(&dev->kobj, name);
-		per_cpu(threshold_banks, i)[bank] = NULL;
-	}
-
 	deallocate_threshold_block(cpu, bank);
 
 free_out:
 	kobject_del(b->kobj);
 	kobject_put(b->kobj);
-	free_cpumask_var(b->cpus);
 	kfree(b);
 	per_cpu(threshold_banks, cpu)[bank] = NULL;
 }
-- 
1.7.9.3.362.g71319

[PATCH 3/9] x86, MCE, AMD: Remove local_allocate_... wrapper

[Borislav Petkov]

From: Borislav Petkov <borislav.petkov@amd.com>

It is unneeded now so drop it.

Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
---
 arch/x86/kernel/cpu/mcheck/mce_amd.c |   11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)

diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c
index 2a5dd30f9969..7fd02cac962b 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c
@@ -61,6 +61,7 @@ struct threshold_bank {
 	struct kobject		*kobj;
 	struct threshold_block	*blocks;
 };
+
 static DEFINE_PER_CPU(struct threshold_bank * [NR_BANKS], threshold_banks);
 
 static unsigned char shared_bank[NR_BANKS] = {
@@ -545,13 +546,6 @@ out_free:
 	return err;
 }
 
-static __cpuinit long
-local_allocate_threshold_blocks(int cpu, unsigned int bank)
-{
-	return allocate_threshold_blocks(cpu, bank, 0,
-					 MSR_IA32_MC0_MISC + bank * 4);
-}
-
 static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank)
 {
 	struct device *dev = per_cpu(mce_device, cpu);
@@ -575,7 +569,8 @@ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank)
 
 	per_cpu(threshold_banks, cpu)[bank] = b;
 
-	err = local_allocate_threshold_blocks(cpu, bank);
+	err = allocate_threshold_blocks(cpu, bank, 0,
+					MSR_IA32_MC0_MISC + bank * 4);
 	if (!err)
 		goto out;
 
-- 
1.7.9.3.362.g71319

[PATCH 4/9] x86, MCE, AMD: Move shared bank to node descriptor

[Borislav Petkov]

From: Borislav Petkov <borislav.petkov@amd.com>

Well, instead of having a real bank 4 on the BSP of each node and
symlinks on the remaining cores, we push it up into the amd_northbridge
descriptor which now contains a pointer to the northbridge bank 4
because the bank is one per northbridge and, as such, belongs in the NB
descriptor anyway.

Each time we hotplug CPUs, we use the northbridge pointer to copy the
shared bank into the per-CPU array of threshold_banks pointers, or
destroy it when the last CPU on the node goes offline, or create it when
the first comes online.

Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
---
 arch/x86/include/asm/amd_nb.h        |   21 +++++++
 arch/x86/kernel/cpu/mcheck/mce_amd.c |  107 +++++++++++++++++++++++++++-------
 2 files changed, 108 insertions(+), 20 deletions(-)

diff --git a/arch/x86/include/asm/amd_nb.h b/arch/x86/include/asm/amd_nb.h
index 49ad773f4b9f..b3341e9cd8fd 100644
--- a/arch/x86/include/asm/amd_nb.h
+++ b/arch/x86/include/asm/amd_nb.h
@@ -26,10 +26,31 @@ struct amd_l3_cache {
 	u8	 subcaches[4];
 };
 
+struct threshold_block {
+	unsigned int		block;
+	unsigned int		bank;
+	unsigned int		cpu;
+	u32			address;
+	u16			interrupt_enable;
+	bool			interrupt_capable;
+	u16			threshold_limit;
+	struct kobject		kobj;
+	struct list_head	miscj;
+};
+
+struct threshold_bank {
+	struct kobject		*kobj;
+	struct threshold_block	*blocks;
+
+	/* initialized to the number of CPUs on the node sharing this bank */
+	atomic_t		cpus;
+};
+
 struct amd_northbridge {
 	struct pci_dev *misc;
 	struct pci_dev *link;
 	struct amd_l3_cache l3_cache;
+	struct threshold_bank *bank4;
 };
 
 struct amd_northbridge_info {
diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c
index 7fd02cac962b..d67c9e56d609 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c
@@ -25,6 +25,7 @@
 #include <linux/cpu.h>
 #include <linux/smp.h>
 
+#include <asm/amd_nb.h>
 #include <asm/apic.h>
 #include <asm/idle.h>
 #include <asm/mce.h>
@@ -45,23 +46,6 @@
 #define MASK_BLKPTR_LO    0xFF000000
 #define MCG_XBLK_ADDR     0xC0000400
 
-struct threshold_block {
-	unsigned int		block;
-	unsigned int		bank;
-	unsigned int		cpu;
-	u32			address;
-	u16			interrupt_enable;
-	bool			interrupt_capable;
-	u16			threshold_limit;
-	struct kobject		kobj;
-	struct list_head	miscj;
-};
-
-struct threshold_bank {
-	struct kobject		*kobj;
-	struct threshold_block	*blocks;
-};
-
 static DEFINE_PER_CPU(struct threshold_bank * [NR_BANKS], threshold_banks);
 
 static unsigned char shared_bank[NR_BANKS] = {
@@ -546,15 +530,62 @@ out_free:
 	return err;
 }
 
+static __cpuinit int __threshold_add_blocks(struct threshold_bank *b)
+{
+	struct list_head *head = &b->blocks->miscj;
+	struct threshold_block *pos = NULL;
+	struct threshold_block *tmp = NULL;
+	int err = 0;
+
+	err = kobject_add(&b->blocks->kobj, b->kobj, b->blocks->kobj.name);
+	if (err)
+		return err;
+
+	list_for_each_entry_safe(pos, tmp, head, miscj) {
+
+		err = kobject_add(&pos->kobj, b->kobj, pos->kobj.name);
+		if (err) {
+			list_for_each_entry_safe_reverse(pos, tmp, head, miscj)
+				kobject_del(&pos->kobj);
+
+			return err;
+		}
+	}
+	return err;
+}
+
 static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank)
 {
 	struct device *dev = per_cpu(mce_device, cpu);
+	struct amd_northbridge *nb = NULL;
 	struct threshold_bank *b = NULL;
 	char name[32];
 	int err = 0;
 
 	sprintf(name, "threshold_bank%i", bank);
 
+	if (shared_bank[bank]) {
+
+		nb = node_to_amd_nb(amd_get_nb_id(cpu));
+		WARN_ON(!nb);
+
+		/* threshold descriptor already initialized on this node? */
+		if (nb->bank4) {
+			/* yes, use it */
+			b = nb->bank4;
+			err = kobject_add(b->kobj, &dev->kobj, name);
+			if (err)
+				goto out;
+
+			per_cpu(threshold_banks, cpu)[bank] = b;
+			atomic_inc(&b->cpus);
+
+			err = __threshold_add_blocks(b);
+
+			goto out;
+		}
+	}
+
 	b = kzalloc(sizeof(struct threshold_bank), GFP_KERNEL);
 	if (!b) {
 		err = -ENOMEM;
@@ -569,15 +600,23 @@ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank)
 
 	per_cpu(threshold_banks, cpu)[bank] = b;
 
+	if (shared_bank[bank]) {
+		atomic_set(&b->cpus, 1);
+
+		/* nb is already initialized, see above */
+		WARN_ON(nb->bank4);
+		nb->bank4 = b;
+	}
+
 	err = allocate_threshold_blocks(cpu, bank, 0,
 					MSR_IA32_MC0_MISC + bank * 4);
 	if (!err)
 		goto out;
 
-out_free:
-	per_cpu(threshold_banks, cpu)[bank] = NULL;
+ out_free:
 	kfree(b);
-out:
+
+ out:
 	return err;
 }
 
@@ -618,16 +657,44 @@ static void deallocate_threshold_block(unsigned int cpu,
 	per_cpu(threshold_banks, cpu)[bank]->blocks = NULL;
 }
 
+static void __threshold_remove_blocks(struct threshold_bank *b)
+{
+	struct threshold_block *pos = NULL;
+	struct threshold_block *tmp = NULL;
+
+	kobject_del(b->kobj);
+
+	list_for_each_entry_safe(pos, tmp, &b->blocks->miscj, miscj)
+		kobject_del(&pos->kobj);
+}
+
 static void threshold_remove_bank(unsigned int cpu, int bank)
 {
+	struct amd_northbridge *nb;
 	struct threshold_bank *b;
 
 	b = per_cpu(threshold_banks, cpu)[bank];
 	if (!b)
 		return;
+
 	if (!b->blocks)
 		goto free_out;
 
+	if (shared_bank[bank]) {
+		if (!atomic_dec_and_test(&b->cpus)) {
+			__threshold_remove_blocks(b);
+			per_cpu(threshold_banks, cpu)[bank] = NULL;
+			return;
+		} else {
+			/*
+			 * the last CPU on this node using the shared bank is
+			 * going away, remove that bank now.
+			 */
+			nb = node_to_amd_nb(amd_get_nb_id(cpu));
+			nb->bank4 = NULL;
+		}
+	}
+
 	deallocate_threshold_block(cpu, bank);
 
 free_out:
-- 
1.7.9.3.362.g71319

[PATCH 5/9] x86, MCE, AMD: Print decimal thresholding values

[Borislav Petkov]

From: Borislav Petkov <borislav.petkov@amd.com>

If one sets the threshold limit, say to 25:

$ echo 25 > machinecheck0/threshold_bank4/misc0/threshold_limit

and then reads it back again, it gives

$ cat machinecheck0/threshold_bank4/misc0/threshold_limit
19

which is actually 0x19 but we don't know that.

Make all output decimal.

Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
---
 arch/x86/kernel/cpu/mcheck/mce_amd.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c
index d67c9e56d609..0b1bb0e15881 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c
@@ -308,7 +308,7 @@ struct threshold_attr {
 #define SHOW_FIELDS(name)						\
 static ssize_t show_ ## name(struct threshold_block *b, char *buf)	\
 {									\
-	return sprintf(buf, "%lx\n", (unsigned long) b->name);		\
+	return sprintf(buf, "%lu\n", (unsigned long) b->name);		\
 }
 SHOW_FIELDS(interrupt_enable)
 SHOW_FIELDS(threshold_limit)
@@ -379,7 +379,7 @@ static ssize_t show_error_count(struct threshold_block *b, char *buf)
 	struct threshold_block_cross_cpu tbcc = { .tb = b, };
 
 	smp_call_function_single(b->cpu, local_error_count_handler, &tbcc, 1);
-	return sprintf(buf, "%lx\n", tbcc.retval);
+	return sprintf(buf, "%lu\n", tbcc.retval);
 }
 
 static ssize_t store_error_count(struct threshold_block *b,
-- 
1.7.9.3.362.g71319

[PATCH 6/9] x86, MCE, AMD: Cleanup reading of error_count

[Borislav Petkov]

From: Borislav Petkov <borislav.petkov@amd.com>

We have rdmsr_on_cpu() now so remove locally defined solution in favor
of the generic one.

No functionality change.

Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
---
 arch/x86/kernel/cpu/mcheck/mce_amd.c |   23 +++++------------------
 1 file changed, 5 insertions(+), 18 deletions(-)

diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c
index 0b1bb0e15881..a7204ef37223 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c
@@ -359,27 +359,14 @@ store_threshold_limit(struct threshold_block *b, const char *buf, size_t size)
 	return size;
 }
 
-struct threshold_block_cross_cpu {
-	struct threshold_block	*tb;
-	long			retval;
-};
-
-static void local_error_count_handler(void *_tbcc)
-{
-	struct threshold_block_cross_cpu *tbcc = _tbcc;
-	struct threshold_block *b = tbcc->tb;
-	u32 low, high;
-
-	rdmsr(b->address, low, high);
-	tbcc->retval = (high & 0xFFF) - (THRESHOLD_MAX - b->threshold_limit);
-}
-
 static ssize_t show_error_count(struct threshold_block *b, char *buf)
 {
-	struct threshold_block_cross_cpu tbcc = { .tb = b, };
+	u32 lo, hi;
+
+	rdmsr_on_cpu(b->cpu, b->address, &lo, &hi);
 
-	smp_call_function_single(b->cpu, local_error_count_handler, &tbcc, 1);
-	return sprintf(buf, "%lu\n", tbcc.retval);
+	return sprintf(buf, "%u\n", ((hi & THRESHOLD_MAX) -
+				     (THRESHOLD_MAX - b->threshold_limit)));
 }
 
 static ssize_t store_error_count(struct threshold_block *b,
-- 
1.7.9.3.362.g71319

[PATCH 7/9] x86, MCE, AMD: Make error_count read only

[Borislav Petkov]

From: Borislav Petkov <borislav.petkov@amd.com>

Until now, writing to error count caused the code to reset the
thresholding bank to the current thresholding limit and start counting
errors from the beginning.

This is misleading and unclear, and can be accomplished by writing the
old thresholding limit into ->threshold_limit.

Make error_count read-only with the functionality to show the current
error count.

Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
---
 arch/x86/kernel/cpu/mcheck/mce_amd.c |   13 ++++---------
 1 file changed, 4 insertions(+), 9 deletions(-)

diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c
index a7204ef37223..e5ed2c7cb4de 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c
@@ -369,14 +369,10 @@ static ssize_t show_error_count(struct threshold_block *b, char *buf)
 				     (THRESHOLD_MAX - b->threshold_limit)));
 }
 
-static ssize_t store_error_count(struct threshold_block *b,
-				 const char *buf, size_t count)
-{
-	struct thresh_restart tr = { .b = b, .reset = 1, .old_limit = 0 };
-
-	smp_call_function_single(b->cpu, threshold_restart_bank, &tr, 1);
-	return 1;
-}
+static struct threshold_attr error_count = {
+	.attr = {.name = __stringify(error_count), .mode = 0444 },
+	.show = show_error_count,
+};
 
 #define RW_ATTR(val)							\
 static struct threshold_attr val = {					\
@@ -387,7 +383,6 @@ static struct threshold_attr val = {					\
 
 RW_ATTR(interrupt_enable);
 RW_ATTR(threshold_limit);
-RW_ATTR(error_count);
 
 static struct attribute *default_attrs[] = {
 	&threshold_limit.attr,
-- 
1.7.9.3.362.g71319

[PATCH 8/9] x86, MCE, AMD: Give proper names to the thresholding banks

[Borislav Petkov]

From: Borislav Petkov <borislav.petkov@amd.com>

Having the banks numbered is ok but having real names which mean
something to the user makes a lot more sense:

/sys/devices/system/machinecheck/machinecheck0/
├── bank0
├── bank1
├── bank2
├── bank3
├── bank4
├── bank5
├── bank6
├── check_interval
├── cmci_disabled
├── combined_unit
│   └── combined_unit
│       ├── error_count
│       └── threshold_limit
├── dont_log_ce
├── execution_unit
│   └── execution_unit
│       ├── error_count
│       └── threshold_limit
├── ignore_ce
├── insn_fetch
│   └── insn_fetch
│       ├── error_count
│       └── threshold_limit
├── load_store
│   └── load_store
│       ├── error_count
│       └── threshold_limit
├── monarch_timeout
├── northbridge
│   ├── dram
│   │   ├── error_count
│   │   ├── interrupt_enable
│   │   └── threshold_limit
│   ├── ht_links
│   │   ├── error_count
│   │   ├── interrupt_enable
│   │   └── threshold_limit
│   └── l3_cache
│       ├── error_count
│       ├── interrupt_enable
│       └── threshold_limit
...

Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
---
 arch/x86/kernel/cpu/mcheck/mce_amd.c |   35 ++++++++++++++++++++++++++++++----
 1 file changed, 31 insertions(+), 4 deletions(-)

diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c
index e5ed2c7cb4de..e20bdf8d7c59 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c
@@ -46,6 +46,15 @@
 #define MASK_BLKPTR_LO    0xFF000000
 #define MCG_XBLK_ADDR     0xC0000400
 
+static const char * const th_names[] = {
+	"load_store",
+	"insn_fetch",
+	"combined_unit",
+	"",
+	"northbridge",
+	"execution_unit",
+};
+
 static DEFINE_PER_CPU(struct threshold_bank * [NR_BANKS], threshold_banks);
 
 static unsigned char shared_bank[NR_BANKS] = {
@@ -68,6 +77,26 @@ struct thresh_restart {
 	u16			old_limit;
 };
 
+static const char * const bank4_names(struct threshold_block *b)
+{
+	switch (b->address) {
+	/* MSR4_MISC0 */
+	case 0x00000413:
+		return "dram";
+
+	case 0xc0000408:
+		return "ht_links";
+
+	case 0xc0000409:
+		return "l3_cache";
+
+	default:
+		WARN(1, "Funny MSR: 0x%08x\n", b->address);
+		return "";
+	}
+};
+
+
 static bool lvt_interrupt_supported(unsigned int bank, u32 msr_high_bits)
 {
 	/*
@@ -481,7 +510,7 @@ static __cpuinit int allocate_threshold_blocks(unsigned int cpu,
 
 	err = kobject_init_and_add(&b->kobj, &threshold_ktype,
 				   per_cpu(threshold_banks, cpu)[bank]->kobj,
-				   "misc%i", block);
+				   (bank == 4 ? bank4_names(b) : th_names[bank]));
 	if (err)
 		goto out_free;
 recurse:
@@ -541,11 +570,9 @@ static __cpuinit int threshold_create_bank(unsigned int cpu, unsigned int bank)
 	struct device *dev = per_cpu(mce_device, cpu);
 	struct amd_northbridge *nb = NULL;
 	struct threshold_bank *b = NULL;
-	char name[32];
+	const char *name = th_names[bank];
 	int err = 0;
 
-	sprintf(name, "threshold_bank%i", bank);
-
 	if (shared_bank[bank]) {
 
 		nb = node_to_amd_nb(amd_get_nb_id(cpu));
-- 
1.7.9.3.362.g71319

[PATCH 9/9] x86, MCE, AMD: Update copyrights and boilerplate

[Borislav Petkov]

From: Borislav Petkov <borislav.petkov@amd.com>

Jacob is doing something else now so add myself as the loser who
provides support.

Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
---
 arch/x86/kernel/cpu/mcheck/mce_amd.c |    6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd.c b/arch/x86/kernel/cpu/mcheck/mce_amd.c
index e20bdf8d7c59..671b95a2ffb5 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd.c
@@ -1,15 +1,17 @@
 /*
- *  (c) 2005, 2006 Advanced Micro Devices, Inc.
+ *  (c) 2005-2012 Advanced Micro Devices, Inc.
  *  Your use of this code is subject to the terms and conditions of the
  *  GNU general public license version 2. See "COPYING" or
  *  http://www.gnu.org/licenses/gpl.html
  *
  *  Written by Jacob Shin - AMD, Inc.
  *
- *  Support : jacob.shin@amd.com
+ *  Support: borislav.petkov@amd.com
  *
  *  April 2006
  *     - added support for AMD Family 0x10 processors
+ *  May 2012
+ *     - major scrubbing
  *
  *  All MC4_MISCi registers are shared between multi-cores
  */
-- 
1.7.9.3.362.g71319