summaryrefslogtreecommitdiff
path: root/drivers
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2017-11-15 10:56:56 -0800
committerLinus Torvalds <torvalds@linux-foundation.org>2017-11-15 10:56:56 -0800
commitc9b012e5f4a1d01dfa8abc6318211a67ba7d5db2 (patch)
tree97b2f1c654fc4333e9e3111f76a26ec5503ee5b9 /drivers
parentb293fca43be544483b6488d33ad4b3ed55881064 (diff)
parent6cfa7cc46b1a7a15d81d5389c99cfca633c12b8e (diff)
Merge tag 'arm64-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Pull arm64 updates from Will Deacon: "The big highlight is support for the Scalable Vector Extension (SVE) which required extensive ABI work to ensure we don't break existing applications by blowing away their signal stack with the rather large new vector context (<= 2 kbit per vector register). There's further work to be done optimising things like exception return, but the ABI is solid now. Much of the line count comes from some new PMU drivers we have, but they're pretty self-contained and I suspect we'll have more of them in future. Plenty of acronym soup here: - initial support for the Scalable Vector Extension (SVE) - improved handling for SError interrupts (required to handle RAS events) - enable GCC support for 128-bit integer types - remove kernel text addresses from backtraces and register dumps - use of WFE to implement long delay()s - ACPI IORT updates from Lorenzo Pieralisi - perf PMU driver for the Statistical Profiling Extension (SPE) - perf PMU driver for Hisilicon's system PMUs - misc cleanups and non-critical fixes" * tag 'arm64-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux: (97 commits) arm64: Make ARMV8_DEPRECATED depend on SYSCTL arm64: Implement __lshrti3 library function arm64: support __int128 on gcc 5+ arm64/sve: Add documentation arm64/sve: Detect SVE and activate runtime support arm64/sve: KVM: Hide SVE from CPU features exposed to guests arm64/sve: KVM: Treat guest SVE use as undefined instruction execution arm64/sve: KVM: Prevent guests from using SVE arm64/sve: Add sysctl to set the default vector length for new processes arm64/sve: Add prctl controls for userspace vector length management arm64/sve: ptrace and ELF coredump support arm64/sve: Preserve SVE registers around EFI runtime service calls arm64/sve: Preserve SVE registers around kernel-mode NEON use arm64/sve: Probe SVE capabilities and usable vector lengths arm64: cpufeature: Move sys_caps_initialised declarations arm64/sve: Backend logic for setting the vector length arm64/sve: Signal handling support arm64/sve: Support vector length resetting for new processes arm64/sve: Core task context handling arm64/sve: Low-level CPU setup ...
Diffstat (limited to 'drivers')
-rw-r--r--drivers/acpi/arm64/gtdt.c2
-rw-r--r--drivers/acpi/arm64/iort.c258
-rw-r--r--drivers/bus/arm-ccn.c1
-rw-r--r--drivers/clocksource/arm_arch_timer.c25
-rw-r--r--drivers/perf/Kconfig15
-rw-r--r--drivers/perf/Makefile2
-rw-r--r--drivers/perf/arm_pmu.c10
-rw-r--r--drivers/perf/arm_pmu_acpi.c3
-rw-r--r--drivers/perf/arm_pmu_platform.c4
-rw-r--r--drivers/perf/arm_spe_pmu.c1249
-rw-r--r--drivers/perf/hisilicon/Makefile1
-rw-r--r--drivers/perf/hisilicon/hisi_uncore_ddrc_pmu.c463
-rw-r--r--drivers/perf/hisilicon/hisi_uncore_hha_pmu.c473
-rw-r--r--drivers/perf/hisilicon/hisi_uncore_l3c_pmu.c463
-rw-r--r--drivers/perf/hisilicon/hisi_uncore_pmu.c447
-rw-r--r--drivers/perf/hisilicon/hisi_uncore_pmu.h102
-rw-r--r--drivers/perf/qcom_l2_pmu.c54
17 files changed, 3500 insertions, 72 deletions
diff --git a/drivers/acpi/arm64/gtdt.c b/drivers/acpi/arm64/gtdt.c
index 597a737d538f..92f9edf9d11e 100644
--- a/drivers/acpi/arm64/gtdt.c
+++ b/drivers/acpi/arm64/gtdt.c
@@ -199,7 +199,7 @@ static int __init gtdt_parse_timer_block(struct acpi_gtdt_timer_block *block,
struct acpi_gtdt_timer_entry *gtdt_frame;
if (!block->timer_count) {
- pr_err(FW_BUG "GT block present, but frame count is zero.");
+ pr_err(FW_BUG "GT block present, but frame count is zero.\n");
return -ENODEV;
}
diff --git a/drivers/acpi/arm64/iort.c b/drivers/acpi/arm64/iort.c
index de56394dd161..95255ecfae7c 100644
--- a/drivers/acpi/arm64/iort.c
+++ b/drivers/acpi/arm64/iort.c
@@ -88,8 +88,8 @@ static inline int iort_set_fwnode(struct acpi_iort_node *iort_node,
*
* Returns: fwnode_handle pointer on success, NULL on failure
*/
-static inline
-struct fwnode_handle *iort_get_fwnode(struct acpi_iort_node *node)
+static inline struct fwnode_handle *iort_get_fwnode(
+ struct acpi_iort_node *node)
{
struct iort_fwnode *curr;
struct fwnode_handle *fwnode = NULL;
@@ -126,6 +126,31 @@ static inline void iort_delete_fwnode(struct acpi_iort_node *node)
spin_unlock(&iort_fwnode_lock);
}
+/**
+ * iort_get_iort_node() - Retrieve iort_node associated with an fwnode
+ *
+ * @fwnode: fwnode associated with device to be looked-up
+ *
+ * Returns: iort_node pointer on success, NULL on failure
+ */
+static inline struct acpi_iort_node *iort_get_iort_node(
+ struct fwnode_handle *fwnode)
+{
+ struct iort_fwnode *curr;
+ struct acpi_iort_node *iort_node = NULL;
+
+ spin_lock(&iort_fwnode_lock);
+ list_for_each_entry(curr, &iort_fwnode_list, list) {
+ if (curr->fwnode == fwnode) {
+ iort_node = curr->iort_node;
+ break;
+ }
+ }
+ spin_unlock(&iort_fwnode_lock);
+
+ return iort_node;
+}
+
typedef acpi_status (*iort_find_node_callback)
(struct acpi_iort_node *node, void *context);
@@ -306,9 +331,8 @@ static int iort_id_map(struct acpi_iort_id_mapping *map, u8 type, u32 rid_in,
return 0;
}
-static
-struct acpi_iort_node *iort_node_get_id(struct acpi_iort_node *node,
- u32 *id_out, int index)
+static struct acpi_iort_node *iort_node_get_id(struct acpi_iort_node *node,
+ u32 *id_out, int index)
{
struct acpi_iort_node *parent;
struct acpi_iort_id_mapping *map;
@@ -332,7 +356,8 @@ struct acpi_iort_node *iort_node_get_id(struct acpi_iort_node *node,
if (map->flags & ACPI_IORT_ID_SINGLE_MAPPING) {
if (node->type == ACPI_IORT_NODE_NAMED_COMPONENT ||
- node->type == ACPI_IORT_NODE_PCI_ROOT_COMPLEX) {
+ node->type == ACPI_IORT_NODE_PCI_ROOT_COMPLEX ||
+ node->type == ACPI_IORT_NODE_SMMU_V3) {
*id_out = map->output_base;
return parent;
}
@@ -341,6 +366,47 @@ struct acpi_iort_node *iort_node_get_id(struct acpi_iort_node *node,
return NULL;
}
+#if (ACPI_CA_VERSION > 0x20170929)
+static int iort_get_id_mapping_index(struct acpi_iort_node *node)
+{
+ struct acpi_iort_smmu_v3 *smmu;
+
+ switch (node->type) {
+ case ACPI_IORT_NODE_SMMU_V3:
+ /*
+ * SMMUv3 dev ID mapping index was introduced in revision 1
+ * table, not available in revision 0
+ */
+ if (node->revision < 1)
+ return -EINVAL;
+
+ smmu = (struct acpi_iort_smmu_v3 *)node->node_data;
+ /*
+ * ID mapping index is only ignored if all interrupts are
+ * GSIV based
+ */
+ if (smmu->event_gsiv && smmu->pri_gsiv && smmu->gerr_gsiv
+ && smmu->sync_gsiv)
+ return -EINVAL;
+
+ if (smmu->id_mapping_index >= node->mapping_count) {
+ pr_err(FW_BUG "[node %p type %d] ID mapping index overflows valid mappings\n",
+ node, node->type);
+ return -EINVAL;
+ }
+
+ return smmu->id_mapping_index;
+ default:
+ return -EINVAL;
+ }
+}
+#else
+static inline int iort_get_id_mapping_index(struct acpi_iort_node *node)
+{
+ return -EINVAL;
+}
+#endif
+
static struct acpi_iort_node *iort_node_map_id(struct acpi_iort_node *node,
u32 id_in, u32 *id_out,
u8 type_mask)
@@ -350,7 +416,7 @@ static struct acpi_iort_node *iort_node_map_id(struct acpi_iort_node *node,
/* Parse the ID mapping tree to find specified node type */
while (node) {
struct acpi_iort_id_mapping *map;
- int i;
+ int i, index;
if (IORT_TYPE_MASK(node->type) & type_mask) {
if (id_out)
@@ -371,8 +437,19 @@ static struct acpi_iort_node *iort_node_map_id(struct acpi_iort_node *node,
goto fail_map;
}
+ /*
+ * Get the special ID mapping index (if any) and skip its
+ * associated ID map to prevent erroneous multi-stage
+ * IORT ID translations.
+ */
+ index = iort_get_id_mapping_index(node);
+
/* Do the ID translation */
for (i = 0; i < node->mapping_count; i++, map++) {
+ /* if it is special mapping index, skip it */
+ if (i == index)
+ continue;
+
if (!iort_id_map(map, node->type, id, &id))
break;
}
@@ -392,10 +469,9 @@ fail_map:
return NULL;
}
-static
-struct acpi_iort_node *iort_node_map_platform_id(struct acpi_iort_node *node,
- u32 *id_out, u8 type_mask,
- int index)
+static struct acpi_iort_node *iort_node_map_platform_id(
+ struct acpi_iort_node *node, u32 *id_out, u8 type_mask,
+ int index)
{
struct acpi_iort_node *parent;
u32 id;
@@ -424,9 +500,25 @@ static struct acpi_iort_node *iort_find_dev_node(struct device *dev)
{
struct pci_bus *pbus;
- if (!dev_is_pci(dev))
+ if (!dev_is_pci(dev)) {
+ struct acpi_iort_node *node;
+ /*
+ * scan iort_fwnode_list to see if it's an iort platform
+ * device (such as SMMU, PMCG),its iort node already cached
+ * and associated with fwnode when iort platform devices
+ * were initialized.
+ */
+ node = iort_get_iort_node(dev->fwnode);
+ if (node)
+ return node;
+
+ /*
+ * if not, then it should be a platform device defined in
+ * DSDT/SSDT (with Named Component node in IORT)
+ */
return iort_scan_node(ACPI_IORT_NODE_NAMED_COMPONENT,
iort_match_node_callback, dev);
+ }
/* Find a PCI root bus */
pbus = to_pci_dev(dev)->bus;
@@ -466,16 +558,24 @@ u32 iort_msi_map_rid(struct device *dev, u32 req_id)
*/
int iort_pmsi_get_dev_id(struct device *dev, u32 *dev_id)
{
- int i;
+ int i, index;
struct acpi_iort_node *node;
node = iort_find_dev_node(dev);
if (!node)
return -ENODEV;
- for (i = 0; i < node->mapping_count; i++) {
- if (iort_node_map_platform_id(node, dev_id, IORT_MSI_TYPE, i))
+ index = iort_get_id_mapping_index(node);
+ /* if there is a valid index, go get the dev_id directly */
+ if (index >= 0) {
+ if (iort_node_get_id(node, dev_id, index))
return 0;
+ } else {
+ for (i = 0; i < node->mapping_count; i++) {
+ if (iort_node_map_platform_id(node, dev_id,
+ IORT_MSI_TYPE, i))
+ return 0;
+ }
}
return -ENODEV;
@@ -538,6 +638,49 @@ struct irq_domain *iort_get_device_domain(struct device *dev, u32 req_id)
return irq_find_matching_fwnode(handle, DOMAIN_BUS_PCI_MSI);
}
+static void iort_set_device_domain(struct device *dev,
+ struct acpi_iort_node *node)
+{
+ struct acpi_iort_its_group *its;
+ struct acpi_iort_node *msi_parent;
+ struct acpi_iort_id_mapping *map;
+ struct fwnode_handle *iort_fwnode;
+ struct irq_domain *domain;
+ int index;
+
+ index = iort_get_id_mapping_index(node);
+ if (index < 0)
+ return;
+
+ map = ACPI_ADD_PTR(struct acpi_iort_id_mapping, node,
+ node->mapping_offset + index * sizeof(*map));
+
+ /* Firmware bug! */
+ if (!map->output_reference ||
+ !(map->flags & ACPI_IORT_ID_SINGLE_MAPPING)) {
+ pr_err(FW_BUG "[node %p type %d] Invalid MSI mapping\n",
+ node, node->type);
+ return;
+ }
+
+ msi_parent = ACPI_ADD_PTR(struct acpi_iort_node, iort_table,
+ map->output_reference);
+
+ if (!msi_parent || msi_parent->type != ACPI_IORT_NODE_ITS_GROUP)
+ return;
+
+ /* Move to ITS specific data */
+ its = (struct acpi_iort_its_group *)msi_parent->node_data;
+
+ iort_fwnode = iort_find_domain_token(its->identifiers[0]);
+ if (!iort_fwnode)
+ return;
+
+ domain = irq_find_matching_fwnode(iort_fwnode, DOMAIN_BUS_PLATFORM_MSI);
+ if (domain)
+ dev_set_msi_domain(dev, domain);
+}
+
/**
* iort_get_platform_device_domain() - Find MSI domain related to a
* platform device
@@ -623,14 +766,14 @@ static inline bool iort_iommu_driver_enabled(u8 type)
}
#ifdef CONFIG_IOMMU_API
-static inline
-const struct iommu_ops *iort_fwspec_iommu_ops(struct iommu_fwspec *fwspec)
+static inline const struct iommu_ops *iort_fwspec_iommu_ops(
+ struct iommu_fwspec *fwspec)
{
return (fwspec && fwspec->ops) ? fwspec->ops : NULL;
}
-static inline
-int iort_add_device_replay(const struct iommu_ops *ops, struct device *dev)
+static inline int iort_add_device_replay(const struct iommu_ops *ops,
+ struct device *dev)
{
int err = 0;
@@ -640,11 +783,11 @@ int iort_add_device_replay(const struct iommu_ops *ops, struct device *dev)
return err;
}
#else
-static inline
-const struct iommu_ops *iort_fwspec_iommu_ops(struct iommu_fwspec *fwspec)
+static inline const struct iommu_ops *iort_fwspec_iommu_ops(
+ struct iommu_fwspec *fwspec)
{ return NULL; }
-static inline
-int iort_add_device_replay(const struct iommu_ops *ops, struct device *dev)
+static inline int iort_add_device_replay(const struct iommu_ops *ops,
+ struct device *dev)
{ return 0; }
#endif
@@ -968,7 +1111,7 @@ static bool __init arm_smmu_v3_is_coherent(struct acpi_iort_node *node)
return smmu->flags & ACPI_IORT_SMMU_V3_COHACC_OVERRIDE;
}
-#if defined(CONFIG_ACPI_NUMA) && defined(ACPI_IORT_SMMU_V3_PXM_VALID)
+#if defined(CONFIG_ACPI_NUMA)
/*
* set numa proximity domain for smmuv3 device
*/
@@ -1051,34 +1194,34 @@ static bool __init arm_smmu_is_coherent(struct acpi_iort_node *node)
return smmu->flags & ACPI_IORT_SMMU_COHERENT_WALK;
}
-struct iort_iommu_config {
+struct iort_dev_config {
const char *name;
- int (*iommu_init)(struct acpi_iort_node *node);
- bool (*iommu_is_coherent)(struct acpi_iort_node *node);
- int (*iommu_count_resources)(struct acpi_iort_node *node);
- void (*iommu_init_resources)(struct resource *res,
+ int (*dev_init)(struct acpi_iort_node *node);
+ bool (*dev_is_coherent)(struct acpi_iort_node *node);
+ int (*dev_count_resources)(struct acpi_iort_node *node);
+ void (*dev_init_resources)(struct resource *res,
struct acpi_iort_node *node);
- void (*iommu_set_proximity)(struct device *dev,
+ void (*dev_set_proximity)(struct device *dev,
struct acpi_iort_node *node);
};
-static const struct iort_iommu_config iort_arm_smmu_v3_cfg __initconst = {
+static const struct iort_dev_config iort_arm_smmu_v3_cfg __initconst = {
.name = "arm-smmu-v3",
- .iommu_is_coherent = arm_smmu_v3_is_coherent,
- .iommu_count_resources = arm_smmu_v3_count_resources,
- .iommu_init_resources = arm_smmu_v3_init_resources,
- .iommu_set_proximity = arm_smmu_v3_set_proximity,
+ .dev_is_coherent = arm_smmu_v3_is_coherent,
+ .dev_count_resources = arm_smmu_v3_count_resources,
+ .dev_init_resources = arm_smmu_v3_init_resources,
+ .dev_set_proximity = arm_smmu_v3_set_proximity,
};
-static const struct iort_iommu_config iort_arm_smmu_cfg __initconst = {
+static const struct iort_dev_config iort_arm_smmu_cfg __initconst = {
.name = "arm-smmu",
- .iommu_is_coherent = arm_smmu_is_coherent,
- .iommu_count_resources = arm_smmu_count_resources,
- .iommu_init_resources = arm_smmu_init_resources
+ .dev_is_coherent = arm_smmu_is_coherent,
+ .dev_count_resources = arm_smmu_count_resources,
+ .dev_init_resources = arm_smmu_init_resources
};
-static __init
-const struct iort_iommu_config *iort_get_iommu_cfg(struct acpi_iort_node *node)
+static __init const struct iort_dev_config *iort_get_dev_cfg(
+ struct acpi_iort_node *node)
{
switch (node->type) {
case ACPI_IORT_NODE_SMMU_V3:
@@ -1091,31 +1234,28 @@ const struct iort_iommu_config *iort_get_iommu_cfg(struct acpi_iort_node *node)
}
/**
- * iort_add_smmu_platform_device() - Allocate a platform device for SMMU
- * @node: Pointer to SMMU ACPI IORT node
+ * iort_add_platform_device() - Allocate a platform device for IORT node
+ * @node: Pointer to device ACPI IORT node
*
* Returns: 0 on success, <0 failure
*/
-static int __init iort_add_smmu_platform_device(struct acpi_iort_node *node)
+static int __init iort_add_platform_device(struct acpi_iort_node *node,
+ const struct iort_dev_config *ops)
{
struct fwnode_handle *fwnode;
struct platform_device *pdev;
struct resource *r;
enum dev_dma_attr attr;
int ret, count;
- const struct iort_iommu_config *ops = iort_get_iommu_cfg(node);
-
- if (!ops)
- return -ENODEV;
pdev = platform_device_alloc(ops->name, PLATFORM_DEVID_AUTO);
if (!pdev)
return -ENOMEM;
- if (ops->iommu_set_proximity)
- ops->iommu_set_proximity(&pdev->dev, node);
+ if (ops->dev_set_proximity)
+ ops->dev_set_proximity(&pdev->dev, node);
- count = ops->iommu_count_resources(node);
+ count = ops->dev_count_resources(node);
r = kcalloc(count, sizeof(*r), GFP_KERNEL);
if (!r) {
@@ -1123,7 +1263,7 @@ static int __init iort_add_smmu_platform_device(struct acpi_iort_node *node)
goto dev_put;
}
- ops->iommu_init_resources(r, node);
+ ops->dev_init_resources(r, node);
ret = platform_device_add_resources(pdev, r, count);
/*
@@ -1158,12 +1298,14 @@ static int __init iort_add_smmu_platform_device(struct acpi_iort_node *node)
pdev->dev.fwnode = fwnode;
- attr = ops->iommu_is_coherent(node) ?
- DEV_DMA_COHERENT : DEV_DMA_NON_COHERENT;
+ attr = ops->dev_is_coherent && ops->dev_is_coherent(node) ?
+ DEV_DMA_COHERENT : DEV_DMA_NON_COHERENT;
/* Configure DMA for the page table walker */
acpi_dma_configure(&pdev->dev, attr);
+ iort_set_device_domain(&pdev->dev, node);
+
ret = platform_device_add(pdev);
if (ret)
goto dma_deconfigure;
@@ -1216,6 +1358,7 @@ static void __init iort_init_platform_devices(void)
struct fwnode_handle *fwnode;
int i, ret;
bool acs_enabled = false;
+ const struct iort_dev_config *ops;
/*
* iort_table and iort both point to the start of IORT table, but
@@ -1238,16 +1381,15 @@ static void __init iort_init_platform_devices(void)
if (!acs_enabled)
acs_enabled = iort_enable_acs(iort_node);
- if ((iort_node->type == ACPI_IORT_NODE_SMMU) ||
- (iort_node->type == ACPI_IORT_NODE_SMMU_V3)) {
-
+ ops = iort_get_dev_cfg(iort_node);
+ if (ops) {
fwnode = acpi_alloc_fwnode_static();
if (!fwnode)
return;
iort_set_fwnode(iort_node, fwnode);
- ret = iort_add_smmu_platform_device(iort_node);
+ ret = iort_add_platform_device(iort_node, ops);
if (ret) {
iort_delete_fwnode(iort_node);
acpi_free_fwnode_static(fwnode);
diff --git a/drivers/bus/arm-ccn.c b/drivers/bus/arm-ccn.c
index e8c6946fed9d..3063f5312397 100644
--- a/drivers/bus/arm-ccn.c
+++ b/drivers/bus/arm-ccn.c
@@ -1276,6 +1276,7 @@ static int arm_ccn_pmu_init(struct arm_ccn *ccn)
/* Perf driver registration */
ccn->dt.pmu = (struct pmu) {
+ .module = THIS_MODULE,
.attr_groups = arm_ccn_pmu_attr_groups,
.task_ctx_nr = perf_invalid_context,
.event_init = arm_ccn_pmu_event_init,
diff --git a/drivers/clocksource/arm_arch_timer.c b/drivers/clocksource/arm_arch_timer.c
index 0ecf5beb56ec..538bfa8ba9b4 100644
--- a/drivers/clocksource/arm_arch_timer.c
+++ b/drivers/clocksource/arm_arch_timer.c
@@ -77,6 +77,7 @@ static bool arch_timer_mem_use_virtual;
static bool arch_counter_suspend_stop;
static bool vdso_default = true;
+static cpumask_t evtstrm_available = CPU_MASK_NONE;
static bool evtstrm_enable = IS_ENABLED(CONFIG_ARM_ARCH_TIMER_EVTSTREAM);
static int __init early_evtstrm_cfg(char *buf)
@@ -739,6 +740,7 @@ static void arch_timer_evtstrm_enable(int divider)
#ifdef CONFIG_COMPAT
compat_elf_hwcap |= COMPAT_HWCAP_EVTSTRM;
#endif
+ cpumask_set_cpu(smp_processor_id(), &evtstrm_available);
}
static void arch_timer_configure_evtstream(void)
@@ -863,6 +865,16 @@ u32 arch_timer_get_rate(void)
return arch_timer_rate;
}
+bool arch_timer_evtstrm_available(void)
+{
+ /*
+ * We might get called from a preemptible context. This is fine
+ * because availability of the event stream should be always the same
+ * for a preemptible context and context where we might resume a task.
+ */
+ return cpumask_test_cpu(raw_smp_processor_id(), &evtstrm_available);
+}
+
static u64 arch_counter_get_cntvct_mem(void)
{
u32 vct_lo, vct_hi, tmp_hi;
@@ -928,6 +940,8 @@ static int arch_timer_dying_cpu(unsigned int cpu)
{
struct clock_event_device *clk = this_cpu_ptr(arch_timer_evt);
+ cpumask_clear_cpu(smp_processor_id(), &evtstrm_available);
+
arch_timer_stop(clk);
return 0;
}
@@ -937,10 +951,16 @@ static DEFINE_PER_CPU(unsigned long, saved_cntkctl);
static int arch_timer_cpu_pm_notify(struct notifier_block *self,
unsigned long action, void *hcpu)
{
- if (action == CPU_PM_ENTER)
+ if (action == CPU_PM_ENTER) {
__this_cpu_write(saved_cntkctl, arch_timer_get_cntkctl());
- else if (action == CPU_PM_ENTER_FAILED || action == CPU_PM_EXIT)
+
+ cpumask_clear_cpu(smp_processor_id(), &evtstrm_available);
+ } else if (action == CPU_PM_ENTER_FAILED || action == CPU_PM_EXIT) {
arch_timer_set_cntkctl(__this_cpu_read(saved_cntkctl));
+
+ if (elf_hwcap & HWCAP_EVTSTRM)
+ cpumask_set_cpu(smp_processor_id(), &evtstrm_available);
+ }
return NOTIFY_OK;
}
@@ -1016,7 +1036,6 @@ static int __init arch_timer_register(void)
if (err)
goto out_unreg_notify;
-
/* Register and immediately configure the timer on the boot CPU */
err = cpuhp_setup_state(CPUHP_AP_ARM_ARCH_TIMER_STARTING,
"clockevents/arm/arch_timer:starting",
diff --git a/drivers/perf/Kconfig b/drivers/perf/Kconfig
index e5197ffb7422..b8f44b068fc6 100644
--- a/drivers/perf/Kconfig
+++ b/drivers/perf/Kconfig
@@ -17,6 +17,13 @@ config ARM_PMU_ACPI
depends on ARM_PMU && ACPI
def_bool y
+config HISI_PMU
+ bool "HiSilicon SoC PMU"
+ depends on ARM64 && ACPI
+ help
+ Support for HiSilicon SoC uncore performance monitoring
+ unit (PMU), such as: L3C, HHA and DDRC.
+
config QCOM_L2_PMU
bool "Qualcomm Technologies L2-cache PMU"
depends on ARCH_QCOM && ARM64 && ACPI
@@ -43,4 +50,12 @@ config XGENE_PMU
help
Say y if you want to use APM X-Gene SoC performance monitors.
+config ARM_SPE_PMU
+ tristate "Enable support for the ARMv8.2 Statistical Profiling Extension"
+ depends on PERF_EVENTS && ARM64
+ help
+ Enable perf support for the ARMv8.2 Statistical Profiling
+ Extension, which provides periodic sampling of operations in
+ the CPU pipeline and reports this via the perf AUX interface.
+
endmenu
diff --git a/drivers/perf/Makefile b/drivers/perf/Makefile
index 9402dc8ff22c..710a0135bd61 100644
--- a/drivers/perf/Makefile
+++ b/drivers/perf/Makefile
@@ -1,6 +1,8 @@
# SPDX-License-Identifier: GPL-2.0
obj-$(CONFIG_ARM_PMU) += arm_pmu.o arm_pmu_platform.o
obj-$(CONFIG_ARM_PMU_ACPI) += arm_pmu_acpi.o
+obj-$(CONFIG_HISI_PMU) += hisilicon/
obj-$(CONFIG_QCOM_L2_PMU) += qcom_l2_pmu.o
obj-$(CONFIG_QCOM_L3_PMU) += qcom_l3_pmu.o
obj-$(CONFIG_XGENE_PMU) += xgene_pmu.o
+obj-$(CONFIG_ARM_SPE_PMU) += arm_spe_pmu.o
diff --git a/drivers/perf/arm_pmu.c b/drivers/perf/arm_pmu.c
index d14fc2e67f93..7bc5eee96b31 100644
--- a/drivers/perf/arm_pmu.c
+++ b/drivers/perf/arm_pmu.c
@@ -539,7 +539,7 @@ void armpmu_free_irq(struct arm_pmu *armpmu, int cpu)
if (!cpumask_test_and_clear_cpu(cpu, &armpmu->active_irqs))
return;
- if (irq_is_percpu(irq)) {
+ if (irq_is_percpu_devid(irq)) {
free_percpu_irq(irq, &hw_events->percpu_pmu);
cpumask_clear(&armpmu->active_irqs);
return;
@@ -565,10 +565,10 @@ int armpmu_request_irq(struct arm_pmu *armpmu, int cpu)
if (!irq)
return 0;
- if (irq_is_percpu(irq) && cpumask_empty(&armpmu->active_irqs)) {
+ if (irq_is_percpu_devid(irq) && cpumask_empty(&armpmu->active_irqs)) {
err = request_percpu_irq(irq, handler, "arm-pmu",
&hw_events->percpu_pmu);
- } else if (irq_is_percpu(irq)) {
+ } else if (irq_is_percpu_devid(irq)) {
int other_cpu = cpumask_first(&armpmu->active_irqs);
int other_irq = per_cpu(hw_events->irq, other_cpu);
@@ -649,7 +649,7 @@ static int arm_perf_starting_cpu(unsigned int cpu, struct hlist_node *node)
irq = armpmu_get_cpu_irq(pmu, cpu);
if (irq) {
- if (irq_is_percpu(irq)) {
+ if (irq_is_percpu_devid(irq)) {
enable_percpu_irq(irq, IRQ_TYPE_NONE);
return 0;
}
@@ -667,7 +667,7 @@ static int arm_perf_teardown_cpu(unsigned int cpu, struct hlist_node *node)
return 0;
irq = armpmu_get_cpu_irq(pmu, cpu);
- if (irq && irq_is_percpu(irq))
+ if (irq && irq_is_percpu_devid(irq))
disable_percpu_irq(irq);
return 0;
diff --git a/drivers/perf/arm_pmu_acpi.c b/drivers/perf/arm_pmu_acpi.c
index 3303dd8d8eb5..705f1a390e31 100644
--- a/drivers/perf/arm_pmu_acpi.c
+++ b/drivers/perf/arm_pmu_acpi.c
@@ -193,9 +193,6 @@ int arm_pmu_acpi_probe(armpmu_init_fn init_fn)
int pmu_idx = 0;
int cpu, ret;
- if (acpi_disabled)
- return 0;
-
/*
* Initialise and register the set of PMUs which we know about right
* now. Ideally we'd do this in arm_pmu_acpi_cpu_starting() so that we
diff --git a/drivers/perf/arm_pmu_platform.c b/drivers/perf/arm_pmu_platform.c
index 4428852e1da1..91b224eced18 100644
--- a/drivers/perf/arm_pmu_platform.c
+++ b/drivers/perf/arm_pmu_platform.c
@@ -127,7 +127,7 @@ static int pmu_parse_irqs(struct arm_pmu *pmu)
if (num_irqs == 1) {
int irq = platform_get_irq(pdev, 0);
- if (irq && irq_is_percpu(irq))
+ if (irq && irq_is_percpu_devid(irq))
return pmu_parse_percpu_irq(pmu, irq);
}
@@ -150,7 +150,7 @@ static int pmu_parse_irqs(struct arm_pmu *pmu)
if (WARN_ON(irq <= 0))
continue;
- if (irq_is_percpu(irq)) {
+ if (irq_is_percpu_devid(irq)) {
pr_warn("multiple PPIs or mismatched SPI/PPI detected\n");
return -EINVAL;
}
diff --git a/drivers/perf/arm_spe_pmu.c b/drivers/perf/arm_spe_pmu.c
new file mode 100644
index 000000000000..8ce262fc2561
--- /dev/null
+++ b/drivers/perf/arm_spe_pmu.c
@@ -0,0 +1,1249 @@
+/*
+ * Perf support for the Statistical Profiling Extension, introduced as
+ * part of ARMv8.2.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ *
+ * Copyright (C) 2016 ARM Limited
+ *
+ * Author: Will Deacon <will.deacon@arm.com>
+ */
+
+#define PMUNAME "arm_spe"
+#define DRVNAME PMUNAME "_pmu"
+#define pr_fmt(fmt) DRVNAME ": " fmt
+
+#include <linux/cpuhotplug.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/module.h>
+#include <linux/of_address.h>
+#include <linux/of_device.h>
+#include <linux/perf_event.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#include <asm/sysreg.h>
+
+#define ARM_SPE_BUF_PAD_BYTE 0
+
+struct arm_spe_pmu_buf {
+ int nr_pages;
+ bool snapshot;
+ void *base;
+};
+
+struct arm_spe_pmu {
+ struct pmu pmu;
+ struct platform_device *pdev;
+ cpumask_t supported_cpus;
+ struct hlist_node hotplug_node;
+
+ int irq; /* PPI */
+
+ u16 min_period;
+ u16 counter_sz;
+
+#define SPE_PMU_FEAT_FILT_EVT (1UL << 0)
+#define SPE_PMU_FEAT_FILT_TYP (1UL << 1)
+#define SPE_PMU_FEAT_FILT_LAT (1UL << 2)
+#define SPE_PMU_FEAT_ARCH_INST (1UL << 3)
+#define SPE_PMU_FEAT_LDS (1UL << 4)
+#define SPE_PMU_FEAT_ERND (1UL << 5)
+#define SPE_PMU_FEAT_DEV_PROBED (1UL << 63)
+ u64 features;
+
+ u16 max_record_sz;
+ u16 align;
+ struct perf_output_handle __percpu *handle;
+};
+
+#define to_spe_pmu(p) (container_of(p, struct arm_spe_pmu, pmu))
+
+/* Convert a free-running index from perf into an SPE buffer offset */
+#define PERF_IDX2OFF(idx, buf) ((idx) % ((buf)->nr_pages << PAGE_SHIFT))
+
+/* Keep track of our dynamic hotplug state */
+static enum cpuhp_state arm_spe_pmu_online;
+
+enum arm_spe_pmu_buf_fault_action {
+ SPE_PMU_BUF_FAULT_ACT_SPURIOUS,
+ SPE_PMU_BUF_FAULT_ACT_FATAL,
+ SPE_PMU_BUF_FAULT_ACT_OK,
+};
+
+/* This sysfs gunk was really good fun to write. */
+enum arm_spe_pmu_capabilities {
+ SPE_PMU_CAP_ARCH_INST = 0,
+ SPE_PMU_CAP_ERND,
+ SPE_PMU_CAP_FEAT_MAX,
+ SPE_PMU_CAP_CNT_SZ = SPE_PMU_CAP_FEAT_MAX,
+ SPE_PMU_CAP_MIN_IVAL,
+};
+
+static int arm_spe_pmu_feat_caps[SPE_PMU_CAP_FEAT_MAX] = {
+ [SPE_PMU_CAP_ARCH_INST] = SPE_PMU_FEAT_ARCH_INST,
+ [SPE_PMU_CAP_ERND] = SPE_PMU_FEAT_ERND,
+};
+
+static u32 arm_spe_pmu_cap_get(struct arm_spe_pmu *spe_pmu, int cap)
+{
+ if (cap < SPE_PMU_CAP_FEAT_MAX)
+ return !!(spe_pmu->features & arm_spe_pmu_feat_caps[cap]);
+
+ switch (cap) {
+ case SPE_PMU_CAP_CNT_SZ:
+ return spe_pmu->counter_sz;
+ case SPE_PMU_CAP_MIN_IVAL:
+ return spe_pmu->min_period;
+ default:
+ WARN(1, "unknown cap %d\n", cap);
+ }
+
+ return 0;
+}
+
+static ssize_t arm_spe_pmu_cap_show(struct device *dev,
+ struct device_attribute *attr,
+ char *buf)
+{
+ struct platform_device *pdev = to_platform_device(dev);
+ struct arm_spe_pmu *spe_pmu = platform_get_drvdata(pdev);
+ struct dev_ext_attribute *ea =
+ container_of(attr, struct dev_ext_attribute, attr);
+ int cap = (long)ea->var;
+
+ return snprintf(buf, PAGE_SIZE, "%u\n",
+ arm_spe_pmu_cap_get(spe_pmu, cap));
+}
+
+#define SPE_EXT_ATTR_ENTRY(_name, _func, _var) \
+ &((struct dev_ext_attribute[]) { \
+ { __ATTR(_name, S_IRUGO, _func, NULL), (void *)_var } \
+ })[0].attr.attr
+
+#define SPE_CAP_EXT_ATTR_ENTRY(_name, _var) \
+ SPE_EXT_ATTR_ENTRY(_name, arm_spe_pmu_cap_show, _var)
+
+static struct attribute *arm_spe_pmu_cap_attr[] = {
+ SPE_CAP_EXT_ATTR_ENTRY(arch_inst, SPE_PMU_CAP_ARCH_INST),
+ SPE_CAP_EXT_ATTR_ENTRY(ernd, SPE_PMU_CAP_ERND),
+ SPE_CAP_EXT_ATTR_ENTRY(count_size, SPE_PMU_CAP_CNT_SZ),
+ SPE_CAP_EXT_ATTR_ENTRY(min_interval, SPE_PMU_CAP_MIN_IVAL),
+ NULL,
+};
+
+static struct attribute_group arm_spe_pmu_cap_group = {
+ .name = "caps",
+ .attrs = arm_spe_pmu_cap_attr,
+};
+
+/* User ABI */
+#define ATTR_CFG_FLD_ts_enable_CFG config /* PMSCR_EL1.TS */
+#define ATTR_CFG_FLD_ts_enable_LO 0
+#define ATTR_CFG_FLD_ts_enable_HI 0
+#define ATTR_CFG_FLD_pa_enable_CFG config /* PMSCR_EL1.PA */
+#define ATTR_CFG_FLD_pa_enable_LO 1
+#define ATTR_CFG_FLD_pa_enable_HI 1
+#define ATTR_CFG_FLD_pct_enable_CFG config /* PMSCR_EL1.PCT */
+#define ATTR_CFG_FLD_pct_enable_LO 2
+#define ATTR_CFG_FLD_pct_enable_HI 2
+#define ATTR_CFG_FLD_jitter_CFG config /* PMSIRR_EL1.RND */
+#define ATTR_CFG_FLD_jitter_LO 16
+#define ATTR_CFG_FLD_jitter_HI 16
+#define ATTR_CFG_FLD_branch_filter_CFG config /* PMSFCR_EL1.B */
+#define ATTR_CFG_FLD_branch_filter_LO 32
+#define ATTR_CFG_FLD_branch_filter_HI 32
+#define ATTR_CFG_FLD_load_filter_CFG config /* PMSFCR_EL1.LD */
+#define ATTR_CFG_FLD_load_filter_LO 33
+#define ATTR_CFG_FLD_load_filter_HI 33
+#define ATTR_CFG_FLD_store_filter_CFG config /* PMSFCR_EL1.ST */
+#define ATTR_CFG_FLD_store_filter_LO 34
+#define ATTR_CFG_FLD_store_filter_HI 34
+
+#define ATTR_CFG_FLD_event_filter_CFG config1 /* PMSEVFR_EL1 */
+#define ATTR_CFG_FLD_event_filter_LO 0
+#define ATTR_CFG_FLD_event_filter_HI 63
+
+#define ATTR_CFG_FLD_min_latency_CFG config2 /* PMSLATFR_EL1.MINLAT */
+#define ATTR_CFG_FLD_min_latency_LO 0
+#define ATTR_CFG_FLD_min_latency_HI 11
+
+/* Why does everything I do descend into this? */
+#define __GEN_PMU_FORMAT_ATTR(cfg, lo, hi) \
+ (lo) == (hi) ? #cfg ":" #lo "\n" : #cfg ":" #lo "-" #hi
+
+#define _GEN_PMU_FORMAT_ATTR(cfg, lo, hi) \
+ __GEN_PMU_FORMAT_ATTR(cfg, lo, hi)
+
+#define GEN_PMU_FORMAT_ATTR(name) \
+ PMU_FORMAT_ATTR(name, \
+ _GEN_PMU_FORMAT_ATTR(ATTR_CFG_FLD_##name##_CFG, \
+ ATTR_CFG_FLD_##name##_LO, \
+ ATTR_CFG_FLD_##name##_HI))
+
+#define _ATTR_CFG_GET_FLD(attr, cfg, lo, hi) \
+ ((((attr)->cfg) >> lo) & GENMASK(hi - lo, 0))
+
+#define ATTR_CFG_GET_FLD(attr, name) \
+ _ATTR_CFG_GET_FLD(attr, \
+ ATTR_CFG_FLD_##name##_CFG, \
+ ATTR_CFG_FLD_##name##_LO, \
+ ATTR_CFG_FLD_##name##_HI)
+
+GEN_PMU_FORMAT_ATTR(ts_enable);
+GEN_PMU_FORMAT_ATTR(pa_enable);
+GEN_PMU_FORMAT_ATTR(pct_enable);
+GEN_PMU_FORMAT_ATTR(jitter);
+GEN_PMU_FORMAT_ATTR(branch_filter);
+GEN_PMU_FORMAT_ATTR(load_filter);
+GEN_PMU_FORMAT_ATTR(store_filter);
+GEN_PMU_FORMAT_ATTR(event_filter);
+GEN_PMU_FORMAT_ATTR(min_latency);
+
+static struct attribute *arm_spe_pmu_formats_attr[] = {
+ &format_attr_ts_enable.attr,
+ &format_attr_pa_enable.attr,
+ &format_attr_pct_enable.attr,
+ &format_attr_jitter.attr,
+ &format_attr_branch_filter.attr,
+ &format_attr_load_filter.attr,
+ &format_attr_store_filter.attr,
+ &format_attr_event_filter.attr,
+ &format_attr_min_latency.attr,
+ NULL,
+};
+
+static struct attribute_group arm_spe_pmu_format_group = {
+ .name = "format",
+ .attrs = arm_spe_pmu_formats_attr,
+};
+
+static ssize_t arm_spe_pmu_get_attr_cpumask(struct device *dev,
+ struct device_attribute *attr,
+ char *buf)
+{
+ struct platform_device *pdev = to_platform_device(dev);
+ struct arm_spe_pmu *spe_pmu = platform_get_drvdata(pdev);
+
+ return cpumap_print_to_pagebuf(true, buf, &spe_pmu->supported_cpus);
+}
+static DEVICE_ATTR(cpumask, S_IRUGO, arm_spe_pmu_get_attr_cpumask, NULL);
+
+static struct attribute *arm_spe_pmu_attrs[] = {
+ &dev_attr_cpumask.attr,
+ NULL,
+};
+
+static struct attribute_group arm_spe_pmu_group = {
+ .attrs = arm_spe_pmu_attrs,
+};
+
+static const struct attribute_group *arm_spe_pmu_attr_groups[] = {
+ &arm_spe_pmu_group,
+ &arm_spe_pmu_cap_group,
+ &arm_spe_pmu_format_group,
+ NULL,
+};
+
+/* Convert between user ABI and register values */
+static u64 arm_spe_event_to_pmscr(struct perf_event *event)
+{
+ struct perf_event_attr *attr = &event->attr;
+ u64 reg = 0;
+
+ reg |= ATTR_CFG_GET_FLD(attr, ts_enable) << SYS_PMSCR_EL1_TS_SHIFT;
+ reg |= ATTR_CFG_GET_FLD(attr, pa_enable) << SYS_PMSCR_EL1_PA_SHIFT;
+ reg |= ATTR_CFG_GET_FLD(attr, pct_enable) << SYS_PMSCR_EL1_PCT_SHIFT;
+
+ if (!attr->exclude_user)
+ reg |= BIT(SYS_PMSCR_EL1_E0SPE_SHIFT);
+
+ if (!attr->exclude_kernel)
+ reg |= BIT(SYS_PMSCR_EL1_E1SPE_SHIFT);
+
+ if (IS_ENABLED(CONFIG_PID_IN_CONTEXTIDR) && capable(CAP_SYS_ADMIN))
+ reg |= BIT(SYS_PMSCR_EL1_CX_SHIFT);
+
+ return reg;
+}
+
+static void arm_spe_event_sanitise_period(struct perf_event *event)
+{
+ struct arm_spe_pmu *spe_pmu = to_spe_pmu(event->pmu);
+ u64 period = event->hw.sample_period;
+ u64 max_period = SYS_PMSIRR_EL1_INTERVAL_MASK
+ << SYS_PMSIRR_EL1_INTERVAL_SHIFT;
+
+ if (period < spe_pmu->min_period)
+ period = spe_pmu->min_period;
+ else if (period > max_period)
+ period = max_period;
+ else
+ period &= max_period;
+
+ event->hw.sample_period = period;
+}
+
+static u64 arm_spe_event_to_pmsirr(struct perf_event *event)
+{
+ struct perf_event_attr *attr = &event->attr;
+ u64 reg = 0;
+
+ arm_spe_event_sanitise_period(event);
+
+ reg |= ATTR_CFG_GET_FLD(attr, jitter) << SYS_PMSIRR_EL1_RND_SHIFT;
+ reg |= event->hw.sample_period;
+
+ return reg;
+}
+
+static u64 arm_spe_event_to_pmsfcr(struct perf_event *event)
+{
+ struct perf_event_attr *attr = &event->attr;
+ u64 reg = 0;
+
+ reg |= ATTR_CFG_GET_FLD(attr, load_filter) << SYS_PMSFCR_EL1_LD_SHIFT;
+ reg |= ATTR_CFG_GET_FLD(attr, store_filter) << SYS_PMSFCR_EL1_ST_SHIFT;
+ reg |= ATTR_CFG_GET_FLD(attr, branch_filter) << SYS_PMSFCR_EL1_B_SHIFT;
+
+ if (reg)
+ reg |= BIT(SYS_PMSFCR_EL1_FT_SHIFT);
+
+ if (ATTR_CFG_GET_FLD(attr, event_filter))
+ reg |= BIT(SYS_PMSFCR_EL1_FE_SHIFT);
+
+ if (ATTR_CFG_GET_FLD(attr, min_latency))
+ reg |= BIT(SYS_PMSFCR_EL1_FL_SHIFT);
+
+ return reg;
+}
+
+static u64 arm_spe_event_to_pmsevfr(struct perf_event *event)
+{
+ struct perf_event_attr *attr = &event->attr;
+ return ATTR_CFG_GET_FLD(attr, event_filter);
+}
+
+static u64 arm_spe_event_to_pmslatfr(struct perf_event *event)
+{
+ struct perf_event_attr *attr = &event->attr;
+ return ATTR_CFG_GET_FLD(attr, min_latency)
+ << SYS_PMSLATFR_EL1_MINLAT_SHIFT;
+}
+
+static void arm_spe_pmu_pad_buf(struct perf_output_handle *handle, int len)
+{
+ struct arm_spe_pmu_buf *buf = perf_get_aux(handle);
+ u64 head = PERF_IDX2OFF(handle->head, buf);
+
+ memset(buf->base + head, ARM_SPE_BUF_PAD_BYTE, len);
+ if (!buf->snapshot)
+ perf_aux_output_skip(handle, len);
+}
+
+static u64 arm_spe_pmu_next_snapshot_off(struct perf_output_handle *handle)
+{
+ struct arm_spe_pmu_buf *buf = perf_get_aux(handle);
+ struct arm_spe_pmu *spe_pmu = to_spe_pmu(handle->event->pmu);
+ u64 head = PERF_IDX2OFF(handle->head, buf);
+ u64 limit = buf->nr_pages * PAGE_SIZE;
+
+ /*
+ * The trace format isn't parseable in reverse, so clamp
+ * the limit to half of the buffer size in snapshot mode
+ * so that the worst case is half a buffer of records, as
+ * opposed to a single record.
+ */
+ if (head < limit >> 1)
+ limit >>= 1;
+
+ /*
+ * If we're within max_record_sz of the limit, we must
+ * pad, move the head index and recompute the limit.
+ */
+ if (limit - head < spe_pmu->max_record_sz) {
+ arm_spe_pmu_pad_buf(handle, limit - head);
+ handle->head = PERF_IDX2OFF(limit, buf);
+ limit = ((buf->nr_pages * PAGE_SIZE) >> 1) + handle->head;
+ }
+
+ return limit;
+}
+
+static u64 __arm_spe_pmu_next_off(struct perf_output_handle *handle)
+{
+ struct arm_spe_pmu *spe_pmu = to_spe_pmu(handle->event->pmu);
+ struct arm_spe_pmu_buf *buf = perf_get_aux(handle);
+ const u64 bufsize = buf->nr_pages * PAGE_SIZE;
+ u64 limit = bufsize;
+ u64 head, tail, wakeup;
+
+ /*
+ * The head can be misaligned for two reasons:
+ *
+ * 1. The hardware left PMBPTR pointing to the first byte after
+ * a record when generating a buffer management event.
+ *
+ * 2. We used perf_aux_output_skip to consume handle->size bytes
+ * and CIRC_SPACE was used to compute the size, which always
+ * leaves one entry free.
+ *
+ * Deal with this by padding to the next alignment boundary and
+ * moving the head index. If we run out of buffer space, we'll
+ * reduce handle->size to zero and end up reporting truncation.
+ */
+ head = PERF_IDX2OFF(handle->head, buf);
+ if (!IS_ALIGNED(head, spe_pmu->align)) {
+ unsigned long delta = roundup(head, spe_pmu->align) - head;
+
+ delta = min(delta, handle->size);
+ arm_spe_pmu_pad_buf(handle, delta);
+ head = PERF_IDX2OFF(handle->head, buf);
+ }
+
+ /* If we've run out of free space, then nothing more to do */
+ if (!handle->size)
+ goto no_space;
+
+ /* Compute the tail and wakeup indices now that we've aligned head */
+ tail = PERF_IDX2OFF(handle->head + handle->size, buf);
+ wakeup = PERF_IDX2OFF(handle->wakeup, buf);
+
+ /*
+ * Avoid clobbering unconsumed data. We know we have space, so
+ * if we see head == tail we know that the buffer is empty. If
+ * head > tail, then there's nothing to clobber prior to
+ * wrapping.
+ */
+ if (head < tail)
+ limit = round_down(tail, PAGE_SIZE);
+
+ /*
+ * Wakeup may be arbitrarily far into the future. If it's not in
+ * the current generation, either we'll wrap before hitting it,
+ * or it's in the past and has been handled already.
+ *
+ * If there's a wakeup before we wrap, arrange to be woken up by
+ * the page boundary following it. Keep the tail boundary if
+ * that's lower.
+ */
+ if (handle->wakeup < (handle->head + handle->size) && head <= wakeup)
+ limit = min(limit, round_up(wakeup, PAGE_SIZE));
+
+ if (limit > head)
+ return limit;
+
+ arm_spe_pmu_pad_buf(handle, handle->size);
+no_space:
+ perf_aux_output_flag(handle, PERF_AUX_FLAG_TRUNCATED);
+ perf_aux_output_end(handle, 0);
+ return 0;
+}
+
+static u64 arm_spe_pmu_next_off(struct perf_output_handle *handle)
+{
+ struct arm_spe_pmu_buf *buf = perf_get_aux(handle);
+ struct arm_spe_pmu *spe_pmu = to_spe_pmu(handle->event->pmu);
+ u64 limit = __arm_spe_pmu_next_off(handle);
+ u64 head = PERF_IDX2OFF(handle->head, buf);
+
+ /*
+ * If the head has come too close to the end of the buffer,
+ * then pad to the end and recompute the limit.
+ */
+ if (limit && (limit - head < spe_pmu->max_record_sz)) {
+ arm_spe_pmu_pad_buf(handle, limit - head);
+ limit = __arm_spe_pmu_next_off(handle);
+ }
+
+ return limit;
+}
+
+static void arm_spe_perf_aux_output_begin(struct perf_output_handle *handle,
+ struct perf_event *event)
+{
+ u64 base, limit;
+ struct arm_spe_pmu_buf *buf;
+
+ /* Start a new aux session */
+ buf = perf_aux_output_begin(handle, event);
+ if (!buf) {
+ event->hw.state |= PERF_HES_STOPPED;
+ /*
+ * We still need to clear the limit pointer, since the
+ * profiler might only be disabled by virtue of a fault.
+ */
+ limit = 0;
+ goto out_write_limit;
+ }
+
+ limit = buf->snapshot ? arm_spe_pmu_next_snapshot_off(handle)
+ : arm_spe_pmu_next_off(handle);
+ if (limit)
+ limit |= BIT(SYS_PMBLIMITR_EL1_E_SHIFT);
+
+ limit += (u64)buf->base;
+ base = (u64)buf->base + PERF_IDX2OFF(handle->head, buf);
+ write_sysreg_s(base, SYS_PMBPTR_EL1);
+
+out_write_limit:
+ write_sysreg_s(limit, SYS_PMBLIMITR_EL1);
+}
+
+static void arm_spe_perf_aux_output_end(struct perf_output_handle *handle)
+{
+ struct arm_spe_pmu_buf *buf = perf_get_aux(handle);
+ u64 offset, size;
+
+ offset = read_sysreg_s(SYS_PMBPTR_EL1) - (u64)buf->base;
+ size = offset - PERF_IDX2OFF(handle->head, buf);
+
+ if (buf->snapshot)
+ handle->head = offset;
+
+ perf_aux_output_end(handle, size);
+}
+
+static void arm_spe_pmu_disable_and_drain_local(void)
+{
+ /* Disable profiling at EL0 and EL1 */
+ write_sysreg_s(0, SYS_PMSCR_EL1);
+ isb();
+
+ /* Drain any buffered data */
+ psb_csync();
+ dsb(nsh);
+
+ /* Disable the profiling buffer */
+ write_sysreg_s(0, SYS_PMBLIMITR_EL1);
+ isb();
+}
+
+/* IRQ handling */
+static enum arm_spe_pmu_buf_fault_action
+arm_spe_pmu_buf_get_fault_act(struct perf_output_handle *handle)
+{
+ const char *err_str;
+ u64 pmbsr;
+ enum arm_spe_pmu_buf_fault_action ret;
+
+ /*
+ * Ensure new profiling data is visible to the CPU and any external
+ * aborts have been resolved.
+ */
+ psb_csync();
+ dsb(nsh);
+
+ /* Ensure hardware updates to PMBPTR_EL1 are visible */
+ isb();
+
+ /* Service required? */
+ pmbsr = read_sysreg_s(SYS_PMBSR_EL1);
+ if (!(pmbsr & BIT(SYS_PMBSR_EL1_S_SHIFT)))
+ return SPE_PMU_BUF_FAULT_ACT_SPURIOUS;
+
+ /*
+ * If we've lost data, disable profiling and also set the PARTIAL
+ * flag to indicate that the last record is corrupted.
+ */
+ if (pmbsr & BIT(SYS_PMBSR_EL1_DL_SHIFT))
+ perf_aux_output_flag(handle, PERF_AUX_FLAG_TRUNCATED |
+ PERF_AUX_FLAG_PARTIAL);
+
+ /* Report collisions to userspace so that it can up the period */
+ if (pmbsr & BIT(SYS_PMBSR_EL1_COLL_SHIFT))
+ perf_aux_output_flag(handle, PERF_AUX_FLAG_COLLISION);
+
+ /* We only expect buffer management events */
+ switch (pmbsr & (SYS_PMBSR_EL1_EC_MASK << SYS_PMBSR_EL1_EC_SHIFT)) {
+ case SYS_PMBSR_EL1_EC_BUF:
+ /* Handled below */
+ break;
+ case SYS_PMBSR_EL1_EC_FAULT_S1:
+ case SYS_PMBSR_EL1_EC_FAULT_S2:
+ err_str = "Unexpected buffer fault";
+ goto out_err;
+ default:
+ err_str = "Unknown error code";
+ goto out_err;
+ }
+
+ /* Buffer management event */
+ switch (pmbsr &
+ (SYS_PMBSR_EL1_BUF_BSC_MASK << SYS_PMBSR_EL1_BUF_BSC_SHIFT)) {
+ case SYS_PMBSR_EL1_BUF_BSC_FULL:
+ ret = SPE_PMU_BUF_FAULT_ACT_OK;
+ goto out_stop;
+ default:
+ err_str = "Unknown buffer status code";
+ }
+
+out_err:
+ pr_err_ratelimited("%s on CPU %d [PMBSR=0x%016llx, PMBPTR=0x%016llx, PMBLIMITR=0x%016llx]\n",
+ err_str, smp_processor_id(), pmbsr,
+ read_sysreg_s(SYS_PMBPTR_EL1),
+ read_sysreg_s(SYS_PMBLIMITR_EL1));
+ ret = SPE_PMU_BUF_FAULT_ACT_FATAL;
+
+out_stop:
+ arm_spe_perf_aux_output_end(handle);
+ return ret;
+}
+
+static irqreturn_t arm_spe_pmu_irq_handler(int irq, void *dev)
+{
+ struct perf_output_handle *handle = dev;
+ struct perf_event *event = handle->event;
+ enum arm_spe_pmu_buf_fault_action act;
+
+ if (!perf_get_aux(handle))
+ return IRQ_NONE;
+
+ act = arm_spe_pmu_buf_get_fault_act(handle);
+ if (act == SPE_PMU_BUF_FAULT_ACT_SPURIOUS)
+ return IRQ_NONE;
+
+ /*
+ * Ensure perf callbacks have completed, which may disable the
+ * profiling buffer in response to a TRUNCATION flag.
+ */
+ irq_work_run();
+
+ switch (act) {
+ case SPE_PMU_BUF_FAULT_ACT_FATAL:
+ /*
+ * If a fatal exception occurred then leaving the profiling
+ * buffer enabled is a recipe waiting to happen. Since
+ * fatal faults don't always imply truncation, make sure
+ * that the profiling buffer is disabled explicitly before
+ * clearing the syndrome register.
+ */
+ arm_spe_pmu_disable_and_drain_local();
+ break;
+ case SPE_PMU_BUF_FAULT_ACT_OK:
+ /*
+ * We handled the fault (the buffer was full), so resume
+ * profiling as long as we didn't detect truncation.
+ * PMBPTR might be misaligned, but we'll burn that bridge
+ * when we get to it.
+ */
+ if (!(handle->aux_flags & PERF_AUX_FLAG_TRUNCATED)) {
+ arm_spe_perf_aux_output_begin(handle, event);
+ isb();
+ }
+ break;
+ case SPE_PMU_BUF_FAULT_ACT_SPURIOUS:
+ /* We've seen you before, but GCC has the memory of a sieve. */
+ break;
+ }
+
+ /* The buffer pointers are now sane, so resume profiling. */
+ write_sysreg_s(0, SYS_PMBSR_EL1);
+ return IRQ_HANDLED;
+}
+
+/* Perf callbacks */
+static int arm_spe_pmu_event_init(struct perf_event *event)
+{
+ u64 reg;
+ struct perf_event_attr *attr = &event->attr;
+ struct arm_spe_pmu *spe_pmu = to_spe_pmu(event->pmu);
+
+ /* This is, of course, deeply driver-specific */
+ if (attr->type != event->pmu->type)
+ return -ENOENT;
+
+ if (event->cpu >= 0 &&
+ !cpumask_test_cpu(event->cpu, &spe_pmu->supported_cpus))
+ return -ENOENT;
+
+ if (arm_spe_event_to_pmsevfr(event) & SYS_PMSEVFR_EL1_RES0)
+ return -EOPNOTSUPP;
+
+ if (attr->exclude_idle)
+ return -EOPNOTSUPP;
+
+ /*
+ * Feedback-directed frequency throttling doesn't work when we
+ * have a buffer of samples. We'd need to manually count the
+ * samples in the buffer when it fills up and adjust the event
+ * count to reflect that. Instead, just force the user to specify
+ * a sample period.
+ */
+ if (attr->freq)
+ return -EINVAL;
+
+ reg = arm_spe_event_to_pmsfcr(event);
+ if ((reg & BIT(SYS_PMSFCR_EL1_FE_SHIFT)) &&
+ !(spe_pmu->features & SPE_PMU_FEAT_FILT_EVT))
+ return -EOPNOTSUPP;
+
+ if ((reg & BIT(SYS_PMSFCR_EL1_FT_SHIFT)) &&
+ !(spe_pmu->features & SPE_PMU_FEAT_FILT_TYP))
+ return -EOPNOTSUPP;
+
+ if ((reg & BIT(SYS_PMSFCR_EL1_FL_SHIFT)) &&
+ !(spe_pmu->features & SPE_PMU_FEAT_FILT_LAT))
+ return -EOPNOTSUPP;
+
+ reg = arm_spe_event_to_pmscr(event);
+ if (!capable(CAP_SYS_ADMIN) &&
+ (reg & (BIT(SYS_PMSCR_EL1_PA_SHIFT) |
+ BIT(SYS_PMSCR_EL1_CX_SHIFT) |
+ BIT(SYS_PMSCR_EL1_PCT_SHIFT))))
+ return -EACCES;
+
+ return 0;
+}
+
+static void arm_spe_pmu_start(struct perf_event *event, int flags)
+{
+ u64 reg;
+ struct arm_spe_pmu *spe_pmu = to_spe_pmu(event->pmu);
+ struct hw_perf_event *hwc = &event->hw;
+ struct perf_output_handle *handle = this_cpu_ptr(spe_pmu->handle);
+
+ hwc->state = 0;
+ arm_spe_perf_aux_output_begin(handle, event);
+ if (hwc->state)
+ return;
+
+ reg = arm_spe_event_to_pmsfcr(event);
+ write_sysreg_s(reg, SYS_PMSFCR_EL1);
+
+ reg = arm_spe_event_to_pmsevfr(event);
+ write_sysreg_s(reg, SYS_PMSEVFR_EL1);
+
+ reg = arm_spe_event_to_pmslatfr(event);
+ write_sysreg_s(reg, SYS_PMSLATFR_EL1);
+
+ if (flags & PERF_EF_RELOAD) {
+ reg = arm_spe_event_to_pmsirr(event);
+ write_sysreg_s(reg, SYS_PMSIRR_EL1);
+ isb();
+ reg = local64_read(&hwc->period_left);
+ write_sysreg_s(reg, SYS_PMSICR_EL1);
+ }
+
+ reg = arm_spe_event_to_pmscr(event);
+ isb();
+ write_sysreg_s(reg, SYS_PMSCR_EL1);
+}
+
+static void arm_spe_pmu_stop(struct perf_event *event, int flags)
+{
+ struct arm_spe_pmu *spe_pmu = to_spe_pmu(event->pmu);
+ struct hw_perf_event *hwc = &event->hw;
+ struct perf_output_handle *handle = this_cpu_ptr(spe_pmu->handle);
+
+ /* If we're already stopped, then nothing to do */
+ if (hwc->state & PERF_HES_STOPPED)
+ return;
+
+ /* Stop all trace generation */
+ arm_spe_pmu_disable_and_drain_local();
+
+ if (flags & PERF_EF_UPDATE) {
+ /*
+ * If there's a fault pending then ensure we contain it
+ * to this buffer, since we might be on the context-switch
+ * path.
+ */
+ if (perf_get_aux(handle)) {
+ enum arm_spe_pmu_buf_fault_action act;
+
+ act = arm_spe_pmu_buf_get_fault_act(handle);
+ if (act == SPE_PMU_BUF_FAULT_ACT_SPURIOUS)
+ arm_spe_perf_aux_output_end(handle);
+ else
+ write_sysreg_s(0, SYS_PMBSR_EL1);
+ }
+
+ /*
+ * This may also contain ECOUNT, but nobody else should
+ * be looking at period_left, since we forbid frequency
+ * based sampling.
+ */
+ local64_set(&hwc->period_left, read_sysreg_s(SYS_PMSICR_EL1));
+ hwc->state |= PERF_HES_UPTODATE;
+ }
+
+ hwc->state |= PERF_HES_STOPPED;
+}
+
+static int arm_spe_pmu_add(struct perf_event *event, int flags)
+{
+ int ret = 0;
+ struct arm_spe_pmu *spe_pmu = to_spe_pmu(event->pmu);
+ struct hw_perf_event *hwc = &event->hw;
+ int cpu = event->cpu == -1 ? smp_processor_id() : event->cpu;
+
+ if (!cpumask_test_cpu(cpu, &spe_pmu->supported_cpus))
+ return -ENOENT;
+
+ hwc->state = PERF_HES_UPTODATE | PERF_HES_STOPPED;
+
+ if (flags & PERF_EF_START) {
+ arm_spe_pmu_start(event, PERF_EF_RELOAD);
+ if (hwc->state & PERF_HES_STOPPED)
+ ret = -EINVAL;
+ }
+
+ return ret;
+}
+
+static void arm_spe_pmu_del(struct perf_event *event, int flags)
+{
+ arm_spe_pmu_stop(event, PERF_EF_UPDATE);
+}
+
+static void arm_spe_pmu_read(struct perf_event *event)
+{
+}
+
+static void *arm_spe_pmu_setup_aux(int cpu, void **pages, int nr_pages,
+ bool snapshot)
+{
+ int i;
+ struct page **pglist;
+ struct arm_spe_pmu_buf *buf;
+
+ /* We need at least two pages for this to work. */
+ if (nr_pages < 2)
+ return NULL;
+
+ /*
+ * We require an even number of pages for snapshot mode, so that
+ * we can effectively treat the buffer as consisting of two equal
+ * parts and give userspace a fighting chance of getting some
+ * useful data out of it.
+ */
+ if (!nr_pages || (snapshot && (nr_pages & 1)))
+ return NULL;
+
+ if (cpu == -1)
+ cpu = raw_smp_processor_id();
+
+ buf = kzalloc_node(sizeof(*buf), GFP_KERNEL, cpu_to_node(cpu));
+ if (!buf)
+ return NULL;
+
+ pglist = kcalloc(nr_pages, sizeof(*pglist), GFP_KERNEL);
+ if (!pglist)
+ goto out_free_buf;
+
+ for (i = 0; i < nr_pages; ++i) {
+ struct page *page = virt_to_page(pages[i]);
+
+ if (PagePrivate(page)) {
+ pr_warn("unexpected high-order page for auxbuf!");
+ goto out_free_pglist;
+ }
+
+ pglist[i] = virt_to_page(pages[i]);
+ }
+
+ buf->base = vmap(pglist, nr_pages, VM_MAP, PAGE_KERNEL);
+ if (!buf->base)
+ goto out_free_pglist;
+
+ buf->nr_pages = nr_pages;
+ buf->snapshot = snapshot;
+
+ kfree(pglist);
+ return buf;
+
+out_free_pglist:
+ kfree(pglist);
+out_free_buf:
+ kfree(buf);
+ return NULL;
+}
+
+static void arm_spe_pmu_free_aux(void *aux)
+{
+ struct arm_spe_pmu_buf *buf = aux;
+
+ vunmap(buf->base);
+ kfree(buf);
+}
+
+/* Initialisation and teardown functions */
+static int arm_spe_pmu_perf_init(struct arm_spe_pmu *spe_pmu)
+{
+ static atomic_t pmu_idx = ATOMIC_INIT(-1);
+
+ int idx;
+ char *name;
+ struct device *dev = &spe_pmu->pdev->dev;
+
+ spe_pmu->pmu = (struct pmu) {
+ .module = THIS_MODULE,
+ .capabilities = PERF_PMU_CAP_EXCLUSIVE | PERF_PMU_CAP_ITRACE,
+ .attr_groups = arm_spe_pmu_attr_groups,
+ /*
+ * We hitch a ride on the software context here, so that
+ * we can support per-task profiling (which is not possible
+ * with the invalid context as it doesn't get sched callbacks).
+ * This requires that userspace either uses a dummy event for
+ * perf_event_open, since the aux buffer is not setup until
+ * a subsequent mmap, or creates the profiling event in a
+ * disabled state and explicitly PERF_EVENT_IOC_ENABLEs it
+ * once the buffer has been created.
+ */
+ .task_ctx_nr = perf_sw_context,
+ .event_init = arm_spe_pmu_event_init,
+ .add = arm_spe_pmu_add,
+ .del = arm_spe_pmu_del,
+ .start = arm_spe_pmu_start,
+ .stop = arm_spe_pmu_stop,
+ .read = arm_spe_pmu_read,
+ .setup_aux = arm_spe_pmu_setup_aux,
+ .free_aux = arm_spe_pmu_free_aux,
+ };
+
+ idx = atomic_inc_return(&pmu_idx);
+ name = devm_kasprintf(dev, GFP_KERNEL, "%s_%d", PMUNAME, idx);
+ return perf_pmu_register(&spe_pmu->pmu, name, -1);
+}
+
+static void arm_spe_pmu_perf_destroy(struct arm_spe_pmu *spe_pmu)
+{
+ perf_pmu_unregister(&spe_pmu->pmu);
+}
+
+static void __arm_spe_pmu_dev_probe(void *info)
+{
+ int fld;
+ u64 reg;
+ struct arm_spe_pmu *spe_pmu = info;
+ struct device *dev = &spe_pmu->pdev->dev;
+
+ fld = cpuid_feature_extract_unsigned_field(read_cpuid(ID_AA64DFR0_EL1),
+ ID_AA64DFR0_PMSVER_SHIFT);
+ if (!fld) {
+ dev_err(dev,
+ "unsupported ID_AA64DFR0_EL1.PMSVer [%d] on CPU %d\n",
+ fld, smp_processor_id());
+ return;
+ }
+
+ /* Read PMBIDR first to determine whether or not we have access */
+ reg = read_sysreg_s(SYS_PMBIDR_EL1);
+ if (reg & BIT(SYS_PMBIDR_EL1_P_SHIFT)) {
+ dev_err(dev,
+ "profiling buffer owned by higher exception level\n");
+ return;
+ }
+
+ /* Minimum alignment. If it's out-of-range, then fail the probe */
+ fld = reg >> SYS_PMBIDR_EL1_ALIGN_SHIFT & SYS_PMBIDR_EL1_ALIGN_MASK;
+ spe_pmu->align = 1 << fld;
+ if (spe_pmu->align > SZ_2K) {
+ dev_err(dev, "unsupported PMBIDR.Align [%d] on CPU %d\n",
+ fld, smp_processor_id());
+ return;
+ }
+
+ /* It's now safe to read PMSIDR and figure out what we've got */
+ reg = read_sysreg_s(SYS_PMSIDR_EL1);
+ if (reg & BIT(SYS_PMSIDR_EL1_FE_SHIFT))
+ spe_pmu->features |= SPE_PMU_FEAT_FILT_EVT;
+
+ if (reg & BIT(SYS_PMSIDR_EL1_FT_SHIFT))
+ spe_pmu->features |= SPE_PMU_FEAT_FILT_TYP;
+
+ if (reg & BIT(SYS_PMSIDR_EL1_FL_SHIFT))
+ spe_pmu->features |= SPE_PMU_FEAT_FILT_LAT;
+
+ if (reg & BIT(SYS_PMSIDR_EL1_ARCHINST_SHIFT))
+ spe_pmu->features |= SPE_PMU_FEAT_ARCH_INST;
+
+ if (reg & BIT(SYS_PMSIDR_EL1_LDS_SHIFT))
+ spe_pmu->features |= SPE_PMU_FEAT_LDS;
+
+ if (reg & BIT(SYS_PMSIDR_EL1_ERND_SHIFT))
+ spe_pmu->features |= SPE_PMU_FEAT_ERND;
+
+ /* This field has a spaced out encoding, so just use a look-up */
+ fld = reg >> SYS_PMSIDR_EL1_INTERVAL_SHIFT & SYS_PMSIDR_EL1_INTERVAL_MASK;
+ switch (fld) {
+ case 0:
+ spe_pmu->min_period = 256;
+ break;
+ case 2:
+ spe_pmu->min_period = 512;
+ break;
+ case 3:
+ spe_pmu->min_period = 768;
+ break;
+ case 4:
+ spe_pmu->min_period = 1024;
+ break;
+ case 5:
+ spe_pmu->min_period = 1536;
+ break;
+ case 6:
+ spe_pmu->min_period = 2048;
+ break;
+ case 7:
+ spe_pmu->min_period = 3072;
+ break;
+ default:
+ dev_warn(dev, "unknown PMSIDR_EL1.Interval [%d]; assuming 8\n",
+ fld);
+ /* Fallthrough */
+ case 8:
+ spe_pmu->min_period = 4096;
+ }
+
+ /* Maximum record size. If it's out-of-range, then fail the probe */
+ fld = reg >> SYS_PMSIDR_EL1_MAXSIZE_SHIFT & SYS_PMSIDR_EL1_MAXSIZE_MASK;
+ spe_pmu->max_record_sz = 1 << fld;
+ if (spe_pmu->max_record_sz > SZ_2K || spe_pmu->max_record_sz < 16) {
+ dev_err(dev, "unsupported PMSIDR_EL1.MaxSize [%d] on CPU %d\n",
+ fld, smp_processor_id());
+ return;
+ }
+
+ fld = reg >> SYS_PMSIDR_EL1_COUNTSIZE_SHIFT & SYS_PMSIDR_EL1_COUNTSIZE_MASK;
+ switch (fld) {
+ default:
+ dev_warn(dev, "unknown PMSIDR_EL1.CountSize [%d]; assuming 2\n",
+ fld);
+ /* Fallthrough */
+ case 2:
+ spe_pmu->counter_sz = 12;
+ }
+
+ dev_info(dev,
+ "probed for CPUs %*pbl [max_record_sz %u, align %u, features 0x%llx]\n",
+ cpumask_pr_args(&spe_pmu->supported_cpus),
+ spe_pmu->max_record_sz, spe_pmu->align, spe_pmu->features);
+
+ spe_pmu->features |= SPE_PMU_FEAT_DEV_PROBED;
+ return;
+}
+
+static void __arm_spe_pmu_reset_local(void)
+{
+ /*
+ * This is probably overkill, as we have no idea where we're
+ * draining any buffered data to...
+ */
+ arm_spe_pmu_disable_and_drain_local();
+
+ /* Reset the buffer base pointer */
+ write_sysreg_s(0, SYS_PMBPTR_EL1);
+ isb();
+
+ /* Clear any pending management interrupts */
+ write_sysreg_s(0, SYS_PMBSR_EL1);
+ isb();
+}
+
+static void __arm_spe_pmu_setup_one(void *info)
+{
+ struct arm_spe_pmu *spe_pmu = info;
+
+ __arm_spe_pmu_reset_local();
+ enable_percpu_irq(spe_pmu->irq, IRQ_TYPE_NONE);
+}
+
+static void __arm_spe_pmu_stop_one(void *info)
+{
+ struct arm_spe_pmu *spe_pmu = info;
+
+ disable_percpu_irq(spe_pmu->irq);
+ __arm_spe_pmu_reset_local();
+}
+
+static int arm_spe_pmu_cpu_startup(unsigned int cpu, struct hlist_node *node)
+{
+ struct arm_spe_pmu *spe_pmu;
+
+ spe_pmu = hlist_entry_safe(node, struct arm_spe_pmu, hotplug_node);
+ if (!cpumask_test_cpu(cpu, &spe_pmu->supported_cpus))
+ return 0;
+
+ __arm_spe_pmu_setup_one(spe_pmu);
+ return 0;
+}
+
+static int arm_spe_pmu_cpu_teardown(unsigned int cpu, struct hlist_node *node)
+{
+ struct arm_spe_pmu *spe_pmu;
+
+ spe_pmu = hlist_entry_safe(node, struct arm_spe_pmu, hotplug_node);
+ if (!cpumask_test_cpu(cpu, &spe_pmu->supported_cpus))
+ return 0;
+
+ __arm_spe_pmu_stop_one(spe_pmu);
+ return 0;
+}
+
+static int arm_spe_pmu_dev_init(struct arm_spe_pmu *spe_pmu)
+{
+ int ret;
+ cpumask_t *mask = &spe_pmu->supported_cpus;
+
+ /* Make sure we probe the hardware on a relevant CPU */
+ ret = smp_call_function_any(mask, __arm_spe_pmu_dev_probe, spe_pmu, 1);
+ if (ret || !(spe_pmu->features & SPE_PMU_FEAT_DEV_PROBED))
+ return -ENXIO;
+
+ /* Request our PPIs (note that the IRQ is still disabled) */
+ ret = request_percpu_irq(spe_pmu->irq, arm_spe_pmu_irq_handler, DRVNAME,
+ spe_pmu->handle);
+ if (ret)
+ return ret;
+
+ /*
+ * Register our hotplug notifier now so we don't miss any events.
+ * This will enable the IRQ for any supported CPUs that are already
+ * up.
+ */
+ ret = cpuhp_state_add_instance(arm_spe_pmu_online,
+ &spe_pmu->hotplug_node);
+ if (ret)
+ free_percpu_irq(spe_pmu->irq, spe_pmu->handle);
+
+ return ret;
+}
+
+static void arm_spe_pmu_dev_teardown(struct arm_spe_pmu *spe_pmu)
+{
+ cpuhp_state_remove_instance(arm_spe_pmu_online, &spe_pmu->hotplug_node);
+ free_percpu_irq(spe_pmu->irq, spe_pmu->handle);
+}
+
+/* Driver and device probing */
+static int arm_spe_pmu_irq_probe(struct arm_spe_pmu *spe_pmu)
+{
+ struct platform_device *pdev = spe_pmu->pdev;
+ int irq = platform_get_irq(pdev, 0);
+
+ if (irq < 0) {
+ dev_err(&pdev->dev, "failed to get IRQ (%d)\n", irq);
+ return -ENXIO;
+ }
+
+ if (!irq_is_percpu(irq)) {
+ dev_err(&pdev->dev, "expected PPI but got SPI (%d)\n", irq);
+ return -EINVAL;
+ }
+
+ if (irq_get_percpu_devid_partition(irq, &spe_pmu->supported_cpus)) {
+ dev_err(&pdev->dev, "failed to get PPI partition (%d)\n", irq);
+ return -EINVAL;
+ }
+
+ spe_pmu->irq = irq;
+ return 0;
+}
+
+static const struct of_device_id arm_spe_pmu_of_match[] = {
+ { .compatible = "arm,statistical-profiling-extension-v1", .data = (void *)1 },
+ { /* Sentinel */ },
+};
+
+static int arm_spe_pmu_device_dt_probe(struct platform_device *pdev)
+{
+ int ret;
+ struct arm_spe_pmu *spe_pmu;
+ struct device *dev = &pdev->dev;
+
+ spe_pmu = devm_kzalloc(dev, sizeof(*spe_pmu), GFP_KERNEL);
+ if (!spe_pmu) {
+ dev_err(dev, "failed to allocate spe_pmu\n");
+ return -ENOMEM;
+ }
+
+ spe_pmu->handle = alloc_percpu(typeof(*spe_pmu->handle));
+ if (!spe_pmu->handle)
+ return -ENOMEM;
+
+ spe_pmu->pdev = pdev;
+ platform_set_drvdata(pdev, spe_pmu);
+
+ ret = arm_spe_pmu_irq_probe(spe_pmu);
+ if (ret)
+ goto out_free_handle;
+
+ ret = arm_spe_pmu_dev_init(spe_pmu);
+ if (ret)
+ goto out_free_handle;
+
+ ret = arm_spe_pmu_perf_init(spe_pmu);
+ if (ret)
+ goto out_teardown_dev;
+
+ return 0;
+
+out_teardown_dev:
+ arm_spe_pmu_dev_teardown(spe_pmu);
+out_free_handle:
+ free_percpu(spe_pmu->handle);
+ return ret;
+}
+
+static int arm_spe_pmu_device_remove(struct platform_device *pdev)
+{
+ struct arm_spe_pmu *spe_pmu = platform_get_drvdata(pdev);
+
+ arm_spe_pmu_perf_destroy(spe_pmu);
+ arm_spe_pmu_dev_teardown(spe_pmu);
+ free_percpu(spe_pmu->handle);
+ return 0;
+}
+
+static struct platform_driver arm_spe_pmu_driver = {
+ .driver = {
+ .name = DRVNAME,
+ .of_match_table = of_match_ptr(arm_spe_pmu_of_match),
+ },
+ .probe = arm_spe_pmu_device_dt_probe,
+ .remove = arm_spe_pmu_device_remove,
+};
+
+static int __init arm_spe_pmu_init(void)
+{
+ int ret;
+
+ ret = cpuhp_setup_state_multi(CPUHP_AP_ONLINE_DYN, DRVNAME,
+ arm_spe_pmu_cpu_startup,
+ arm_spe_pmu_cpu_teardown);
+ if (ret < 0)
+ return ret;
+ arm_spe_pmu_online = ret;
+
+ ret = platform_driver_register(&arm_spe_pmu_driver);
+ if (ret)
+ cpuhp_remove_multi_state(arm_spe_pmu_online);
+
+ return ret;
+}
+
+static void __exit arm_spe_pmu_exit(void)
+{
+ platform_driver_unregister(&arm_spe_pmu_driver);
+ cpuhp_remove_multi_state(arm_spe_pmu_online);
+}
+
+module_init(arm_spe_pmu_init);
+module_exit(arm_spe_pmu_exit);
+
+MODULE_DESCRIPTION("Perf driver for the ARMv8.2 Statistical Profiling Extension");
+MODULE_AUTHOR("Will Deacon <will.deacon@arm.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/perf/hisilicon/Makefile b/drivers/perf/hisilicon/Makefile
new file mode 100644
index 000000000000..2621d51ae87a
--- /dev/null
+++ b/drivers/perf/hisilicon/Makefile
@@ -0,0 +1 @@
+obj-$(CONFIG_HISI_PMU) += hisi_uncore_pmu.o hisi_uncore_l3c_pmu.o hisi_uncore_hha_pmu.o hisi_uncore_ddrc_pmu.o
diff --git a/drivers/perf/hisilicon/hisi_uncore_ddrc_pmu.c b/drivers/perf/hisilicon/hisi_uncore_ddrc_pmu.c
new file mode 100644
index 000000000000..1b10ea05a914
--- /dev/null
+++ b/drivers/perf/hisilicon/hisi_uncore_ddrc_pmu.c
@@ -0,0 +1,463 @@
+/*
+ * HiSilicon SoC DDRC uncore Hardware event counters support
+ *
+ * Copyright (C) 2017 Hisilicon Limited
+ * Author: Shaokun Zhang <zhangshaokun@hisilicon.com>
+ * Anurup M <anurup.m@huawei.com>
+ *
+ * This code is based on the uncore PMUs like arm-cci and arm-ccn.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/acpi.h>
+#include <linux/bug.h>
+#include <linux/cpuhotplug.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/list.h>
+#include <linux/platform_device.h>
+#include <linux/smp.h>
+
+#include "hisi_uncore_pmu.h"
+
+/* DDRC register definition */
+#define DDRC_PERF_CTRL 0x010
+#define DDRC_FLUX_WR 0x380
+#define DDRC_FLUX_RD 0x384
+#define DDRC_FLUX_WCMD 0x388
+#define DDRC_FLUX_RCMD 0x38c
+#define DDRC_PRE_CMD 0x3c0
+#define DDRC_ACT_CMD 0x3c4
+#define DDRC_BNK_CHG 0x3c8
+#define DDRC_RNK_CHG 0x3cc
+#define DDRC_EVENT_CTRL 0x6C0
+#define DDRC_INT_MASK 0x6c8
+#define DDRC_INT_STATUS 0x6cc
+#define DDRC_INT_CLEAR 0x6d0
+
+/* DDRC has 8-counters */
+#define DDRC_NR_COUNTERS 0x8
+#define DDRC_PERF_CTRL_EN 0x2
+
+/*
+ * For DDRC PMU, there are eight-events and every event has been mapped
+ * to fixed-purpose counters which register offset is not consistent.
+ * Therefore there is no write event type and we assume that event
+ * code (0 to 7) is equal to counter index in PMU driver.
+ */
+#define GET_DDRC_EVENTID(hwc) (hwc->config_base & 0x7)
+
+static const u32 ddrc_reg_off[] = {
+ DDRC_FLUX_WR, DDRC_FLUX_RD, DDRC_FLUX_WCMD, DDRC_FLUX_RCMD,
+ DDRC_PRE_CMD, DDRC_ACT_CMD, DDRC_BNK_CHG, DDRC_RNK_CHG
+};
+
+/*
+ * Select the counter register offset using the counter index.
+ * In DDRC there are no programmable counter, the count
+ * is readed form the statistics counter register itself.
+ */
+static u32 hisi_ddrc_pmu_get_counter_offset(int cntr_idx)
+{
+ return ddrc_reg_off[cntr_idx];
+}
+
+static u64 hisi_ddrc_pmu_read_counter(struct hisi_pmu *ddrc_pmu,
+ struct hw_perf_event *hwc)
+{
+ /* Use event code as counter index */
+ u32 idx = GET_DDRC_EVENTID(hwc);
+
+ if (!hisi_uncore_pmu_counter_valid(ddrc_pmu, idx)) {
+ dev_err(ddrc_pmu->dev, "Unsupported event index:%d!\n", idx);
+ return 0;
+ }
+
+ return readl(ddrc_pmu->base + hisi_ddrc_pmu_get_counter_offset(idx));
+}
+
+static void hisi_ddrc_pmu_write_counter(struct hisi_pmu *ddrc_pmu,
+ struct hw_perf_event *hwc, u64 val)
+{
+ u32 idx = GET_DDRC_EVENTID(hwc);
+
+ if (!hisi_uncore_pmu_counter_valid(ddrc_pmu, idx)) {
+ dev_err(ddrc_pmu->dev, "Unsupported event index:%d!\n", idx);
+ return;
+ }
+
+ writel((u32)val,
+ ddrc_pmu->base + hisi_ddrc_pmu_get_counter_offset(idx));
+}
+
+/*
+ * For DDRC PMU, event has been mapped to fixed-purpose counter by hardware,
+ * so there is no need to write event type.
+ */
+static void hisi_ddrc_pmu_write_evtype(struct hisi_pmu *hha_pmu, int idx,
+ u32 type)
+{
+}
+
+static void hisi_ddrc_pmu_start_counters(struct hisi_pmu *ddrc_pmu)
+{
+ u32 val;
+
+ /* Set perf_enable in DDRC_PERF_CTRL to start event counting */
+ val = readl(ddrc_pmu->base + DDRC_PERF_CTRL);
+ val |= DDRC_PERF_CTRL_EN;
+ writel(val, ddrc_pmu->base + DDRC_PERF_CTRL);
+}
+
+static void hisi_ddrc_pmu_stop_counters(struct hisi_pmu *ddrc_pmu)
+{
+ u32 val;
+
+ /* Clear perf_enable in DDRC_PERF_CTRL to stop event counting */
+ val = readl(ddrc_pmu->base + DDRC_PERF_CTRL);
+ val &= ~DDRC_PERF_CTRL_EN;
+ writel(val, ddrc_pmu->base + DDRC_PERF_CTRL);
+}
+
+static void hisi_ddrc_pmu_enable_counter(struct hisi_pmu *ddrc_pmu,
+ struct hw_perf_event *hwc)
+{
+ u32 val;
+
+ /* Set counter index(event code) in DDRC_EVENT_CTRL register */
+ val = readl(ddrc_pmu->base + DDRC_EVENT_CTRL);
+ val |= (1 << GET_DDRC_EVENTID(hwc));
+ writel(val, ddrc_pmu->base + DDRC_EVENT_CTRL);
+}
+
+static void hisi_ddrc_pmu_disable_counter(struct hisi_pmu *ddrc_pmu,
+ struct hw_perf_event *hwc)
+{
+ u32 val;
+
+ /* Clear counter index(event code) in DDRC_EVENT_CTRL register */
+ val = readl(ddrc_pmu->base + DDRC_EVENT_CTRL);
+ val &= ~(1 << GET_DDRC_EVENTID(hwc));
+ writel(val, ddrc_pmu->base + DDRC_EVENT_CTRL);
+}
+
+static int hisi_ddrc_pmu_get_event_idx(struct perf_event *event)
+{
+ struct hisi_pmu *ddrc_pmu = to_hisi_pmu(event->pmu);
+ unsigned long *used_mask = ddrc_pmu->pmu_events.used_mask;
+ struct hw_perf_event *hwc = &event->hw;
+ /* For DDRC PMU, we use event code as counter index */
+ int idx = GET_DDRC_EVENTID(hwc);
+
+ if (test_bit(idx, used_mask))
+ return -EAGAIN;
+
+ set_bit(idx, used_mask);
+
+ return idx;
+}
+
+static void hisi_ddrc_pmu_enable_counter_int(struct hisi_pmu *ddrc_pmu,
+ struct hw_perf_event *hwc)
+{
+ u32 val;
+
+ /* Write 0 to enable interrupt */
+ val = readl(ddrc_pmu->base + DDRC_INT_MASK);
+ val &= ~(1 << GET_DDRC_EVENTID(hwc));
+ writel(val, ddrc_pmu->base + DDRC_INT_MASK);
+}
+
+static void hisi_ddrc_pmu_disable_counter_int(struct hisi_pmu *ddrc_pmu,
+ struct hw_perf_event *hwc)
+{
+ u32 val;
+
+ /* Write 1 to mask interrupt */
+ val = readl(ddrc_pmu->base + DDRC_INT_MASK);
+ val |= (1 << GET_DDRC_EVENTID(hwc));
+ writel(val, ddrc_pmu->base + DDRC_INT_MASK);
+}
+
+static irqreturn_t hisi_ddrc_pmu_isr(int irq, void *dev_id)
+{
+ struct hisi_pmu *ddrc_pmu = dev_id;
+ struct perf_event *event;
+ unsigned long overflown;
+ int idx;
+
+ /* Read the DDRC_INT_STATUS register */
+ overflown = readl(ddrc_pmu->base + DDRC_INT_STATUS);
+ if (!overflown)
+ return IRQ_NONE;
+
+ /*
+ * Find the counter index which overflowed if the bit was set
+ * and handle it
+ */
+ for_each_set_bit(idx, &overflown, DDRC_NR_COUNTERS) {
+ /* Write 1 to clear the IRQ status flag */
+ writel((1 << idx), ddrc_pmu->base + DDRC_INT_CLEAR);
+
+ /* Get the corresponding event struct */
+ event = ddrc_pmu->pmu_events.hw_events[idx];
+ if (!event)
+ continue;
+
+ hisi_uncore_pmu_event_update(event);
+ hisi_uncore_pmu_set_event_period(event);
+ }
+
+ return IRQ_HANDLED;
+}
+
+static int hisi_ddrc_pmu_init_irq(struct hisi_pmu *ddrc_pmu,
+ struct platform_device *pdev)
+{
+ int irq, ret;
+
+ /* Read and init IRQ */
+ irq = platform_get_irq(pdev, 0);
+ if (irq < 0) {
+ dev_err(&pdev->dev, "DDRC PMU get irq fail; irq:%d\n", irq);
+ return irq;
+ }
+
+ ret = devm_request_irq(&pdev->dev, irq, hisi_ddrc_pmu_isr,
+ IRQF_NOBALANCING | IRQF_NO_THREAD,
+ dev_name(&pdev->dev), ddrc_pmu);
+ if (ret < 0) {
+ dev_err(&pdev->dev,
+ "Fail to request IRQ:%d ret:%d\n", irq, ret);
+ return ret;
+ }
+
+ ddrc_pmu->irq = irq;
+
+ return 0;
+}
+
+static const struct acpi_device_id hisi_ddrc_pmu_acpi_match[] = {
+ { "HISI0233", },
+ {},
+};
+MODULE_DEVICE_TABLE(acpi, hisi_ddrc_pmu_acpi_match);
+
+static int hisi_ddrc_pmu_init_data(struct platform_device *pdev,
+ struct hisi_pmu *ddrc_pmu)
+{
+ struct resource *res;
+
+ /*
+ * Use the SCCL_ID and DDRC channel ID to identify the
+ * DDRC PMU, while SCCL_ID is in MPIDR[aff2].
+ */
+ if (device_property_read_u32(&pdev->dev, "hisilicon,ch-id",
+ &ddrc_pmu->index_id)) {
+ dev_err(&pdev->dev, "Can not read ddrc channel-id!\n");
+ return -EINVAL;
+ }
+
+ if (device_property_read_u32(&pdev->dev, "hisilicon,scl-id",
+ &ddrc_pmu->sccl_id)) {
+ dev_err(&pdev->dev, "Can not read ddrc sccl-id!\n");
+ return -EINVAL;
+ }
+ /* DDRC PMUs only share the same SCCL */
+ ddrc_pmu->ccl_id = -1;
+
+ res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ ddrc_pmu->base = devm_ioremap_resource(&pdev->dev, res);
+ if (IS_ERR(ddrc_pmu->base)) {
+ dev_err(&pdev->dev, "ioremap failed for ddrc_pmu resource\n");
+ return PTR_ERR(ddrc_pmu->base);
+ }
+
+ return 0;
+}
+
+static struct attribute *hisi_ddrc_pmu_format_attr[] = {
+ HISI_PMU_FORMAT_ATTR(event, "config:0-4"),
+ NULL,
+};
+
+static const struct attribute_group hisi_ddrc_pmu_format_group = {
+ .name = "format",
+ .attrs = hisi_ddrc_pmu_format_attr,
+};
+
+static struct attribute *hisi_ddrc_pmu_events_attr[] = {
+ HISI_PMU_EVENT_ATTR(flux_wr, 0x00),
+ HISI_PMU_EVENT_ATTR(flux_rd, 0x01),
+ HISI_PMU_EVENT_ATTR(flux_wcmd, 0x02),
+ HISI_PMU_EVENT_ATTR(flux_rcmd, 0x03),
+ HISI_PMU_EVENT_ATTR(pre_cmd, 0x04),
+ HISI_PMU_EVENT_ATTR(act_cmd, 0x05),
+ HISI_PMU_EVENT_ATTR(rnk_chg, 0x06),
+ HISI_PMU_EVENT_ATTR(rw_chg, 0x07),
+ NULL,
+};
+
+static const struct attribute_group hisi_ddrc_pmu_events_group = {
+ .name = "events",
+ .attrs = hisi_ddrc_pmu_events_attr,
+};
+
+static DEVICE_ATTR(cpumask, 0444, hisi_cpumask_sysfs_show, NULL);
+
+static struct attribute *hisi_ddrc_pmu_cpumask_attrs[] = {
+ &dev_attr_cpumask.attr,
+ NULL,
+};
+
+static const struct attribute_group hisi_ddrc_pmu_cpumask_attr_group = {
+ .attrs = hisi_ddrc_pmu_cpumask_attrs,
+};
+
+static const struct attribute_group *hisi_ddrc_pmu_attr_groups[] = {
+ &hisi_ddrc_pmu_format_group,
+ &hisi_ddrc_pmu_events_group,
+ &hisi_ddrc_pmu_cpumask_attr_group,
+ NULL,
+};
+
+static const struct hisi_uncore_ops hisi_uncore_ddrc_ops = {
+ .write_evtype = hisi_ddrc_pmu_write_evtype,
+ .get_event_idx = hisi_ddrc_pmu_get_event_idx,
+ .start_counters = hisi_ddrc_pmu_start_counters,
+ .stop_counters = hisi_ddrc_pmu_stop_counters,
+ .enable_counter = hisi_ddrc_pmu_enable_counter,
+ .disable_counter = hisi_ddrc_pmu_disable_counter,
+ .enable_counter_int = hisi_ddrc_pmu_enable_counter_int,
+ .disable_counter_int = hisi_ddrc_pmu_disable_counter_int,
+ .write_counter = hisi_ddrc_pmu_write_counter,
+ .read_counter = hisi_ddrc_pmu_read_counter,
+};
+
+static int hisi_ddrc_pmu_dev_probe(struct platform_device *pdev,
+ struct hisi_pmu *ddrc_pmu)
+{
+ int ret;
+
+ ret = hisi_ddrc_pmu_init_data(pdev, ddrc_pmu);
+ if (ret)
+ return ret;
+
+ ret = hisi_ddrc_pmu_init_irq(ddrc_pmu, pdev);
+ if (ret)
+ return ret;
+
+ ddrc_pmu->num_counters = DDRC_NR_COUNTERS;
+ ddrc_pmu->counter_bits = 32;
+ ddrc_pmu->ops = &hisi_uncore_ddrc_ops;
+ ddrc_pmu->dev = &pdev->dev;
+ ddrc_pmu->on_cpu = -1;
+ ddrc_pmu->check_event = 7;
+
+ return 0;
+}
+
+static int hisi_ddrc_pmu_probe(struct platform_device *pdev)
+{
+ struct hisi_pmu *ddrc_pmu;
+ char *name;
+ int ret;
+
+ ddrc_pmu = devm_kzalloc(&pdev->dev, sizeof(*ddrc_pmu), GFP_KERNEL);
+ if (!ddrc_pmu)
+ return -ENOMEM;
+
+ platform_set_drvdata(pdev, ddrc_pmu);
+
+ ret = hisi_ddrc_pmu_dev_probe(pdev, ddrc_pmu);
+ if (ret)
+ return ret;
+
+ ret = cpuhp_state_add_instance(CPUHP_AP_PERF_ARM_HISI_DDRC_ONLINE,
+ &ddrc_pmu->node);
+ if (ret) {
+ dev_err(&pdev->dev, "Error %d registering hotplug;\n", ret);
+ return ret;
+ }
+
+ name = devm_kasprintf(&pdev->dev, GFP_KERNEL, "hisi_sccl%u_ddrc%u",
+ ddrc_pmu->sccl_id, ddrc_pmu->index_id);
+ ddrc_pmu->pmu = (struct pmu) {
+ .name = name,
+ .task_ctx_nr = perf_invalid_context,
+ .event_init = hisi_uncore_pmu_event_init,
+ .pmu_enable = hisi_uncore_pmu_enable,
+ .pmu_disable = hisi_uncore_pmu_disable,
+ .add = hisi_uncore_pmu_add,
+ .del = hisi_uncore_pmu_del,
+ .start = hisi_uncore_pmu_start,
+ .stop = hisi_uncore_pmu_stop,
+ .read = hisi_uncore_pmu_read,
+ .attr_groups = hisi_ddrc_pmu_attr_groups,
+ };
+
+ ret = perf_pmu_register(&ddrc_pmu->pmu, name, -1);
+ if (ret) {
+ dev_err(ddrc_pmu->dev, "DDRC PMU register failed!\n");
+ cpuhp_state_remove_instance(CPUHP_AP_PERF_ARM_HISI_DDRC_ONLINE,
+ &ddrc_pmu->node);
+ }
+
+ return ret;
+}
+
+static int hisi_ddrc_pmu_remove(struct platform_device *pdev)
+{
+ struct hisi_pmu *ddrc_pmu = platform_get_drvdata(pdev);
+
+ perf_pmu_unregister(&ddrc_pmu->pmu);
+ cpuhp_state_remove_instance(CPUHP_AP_PERF_ARM_HISI_DDRC_ONLINE,
+ &ddrc_pmu->node);
+
+ return 0;
+}
+
+static struct platform_driver hisi_ddrc_pmu_driver = {
+ .driver = {
+ .name = "hisi_ddrc_pmu",
+ .acpi_match_table = ACPI_PTR(hisi_ddrc_pmu_acpi_match),
+ },
+ .probe = hisi_ddrc_pmu_probe,
+ .remove = hisi_ddrc_pmu_remove,
+};
+
+static int __init hisi_ddrc_pmu_module_init(void)
+{
+ int ret;
+
+ ret = cpuhp_setup_state_multi(CPUHP_AP_PERF_ARM_HISI_DDRC_ONLINE,
+ "AP_PERF_ARM_HISI_DDRC_ONLINE",
+ hisi_uncore_pmu_online_cpu,
+ hisi_uncore_pmu_offline_cpu);
+ if (ret) {
+ pr_err("DDRC PMU: setup hotplug, ret = %d\n", ret);
+ return ret;
+ }
+
+ ret = platform_driver_register(&hisi_ddrc_pmu_driver);
+ if (ret)
+ cpuhp_remove_multi_state(CPUHP_AP_PERF_ARM_HISI_DDRC_ONLINE);
+
+ return ret;
+}
+module_init(hisi_ddrc_pmu_module_init);
+
+static void __exit hisi_ddrc_pmu_module_exit(void)
+{
+ platform_driver_unregister(&hisi_ddrc_pmu_driver);
+ cpuhp_remove_multi_state(CPUHP_AP_PERF_ARM_HISI_DDRC_ONLINE);
+
+}
+module_exit(hisi_ddrc_pmu_module_exit);
+
+MODULE_DESCRIPTION("HiSilicon SoC DDRC uncore PMU driver");
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("Shaokun Zhang <zhangshaokun@hisilicon.com>");
+MODULE_AUTHOR("Anurup M <anurup.m@huawei.com>");
diff --git a/drivers/perf/hisilicon/hisi_uncore_hha_pmu.c b/drivers/perf/hisilicon/hisi_uncore_hha_pmu.c
new file mode 100644
index 000000000000..443906e0aff3
--- /dev/null
+++ b/drivers/perf/hisilicon/hisi_uncore_hha_pmu.c
@@ -0,0 +1,473 @@
+/*
+ * HiSilicon SoC HHA uncore Hardware event counters support
+ *
+ * Copyright (C) 2017 Hisilicon Limited
+ * Author: Shaokun Zhang <zhangshaokun@hisilicon.com>
+ * Anurup M <anurup.m@huawei.com>
+ *
+ * This code is based on the uncore PMUs like arm-cci and arm-ccn.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/acpi.h>
+#include <linux/bug.h>
+#include <linux/cpuhotplug.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/list.h>
+#include <linux/platform_device.h>
+#include <linux/smp.h>
+
+#include "hisi_uncore_pmu.h"
+
+/* HHA register definition */
+#define HHA_INT_MASK 0x0804
+#define HHA_INT_STATUS 0x0808
+#define HHA_INT_CLEAR 0x080C
+#define HHA_PERF_CTRL 0x1E00
+#define HHA_EVENT_CTRL 0x1E04
+#define HHA_EVENT_TYPE0 0x1E80
+/*
+ * Each counter is 48-bits and [48:63] are reserved
+ * which are Read-As-Zero and Writes-Ignored.
+ */
+#define HHA_CNT0_LOWER 0x1F00
+
+/* HHA has 16-counters */
+#define HHA_NR_COUNTERS 0x10
+
+#define HHA_PERF_CTRL_EN 0x1
+#define HHA_EVTYPE_NONE 0xff
+
+/*
+ * Select the counter register offset using the counter index
+ * each counter is 48-bits.
+ */
+static u32 hisi_hha_pmu_get_counter_offset(int cntr_idx)
+{
+ return (HHA_CNT0_LOWER + (cntr_idx * 8));
+}
+
+static u64 hisi_hha_pmu_read_counter(struct hisi_pmu *hha_pmu,
+ struct hw_perf_event *hwc)
+{
+ u32 idx = hwc->idx;
+
+ if (!hisi_uncore_pmu_counter_valid(hha_pmu, idx)) {
+ dev_err(hha_pmu->dev, "Unsupported event index:%d!\n", idx);
+ return 0;
+ }
+
+ /* Read 64 bits and like L3C, top 16 bits are RAZ */
+ return readq(hha_pmu->base + hisi_hha_pmu_get_counter_offset(idx));
+}
+
+static void hisi_hha_pmu_write_counter(struct hisi_pmu *hha_pmu,
+ struct hw_perf_event *hwc, u64 val)
+{
+ u32 idx = hwc->idx;
+
+ if (!hisi_uncore_pmu_counter_valid(hha_pmu, idx)) {
+ dev_err(hha_pmu->dev, "Unsupported event index:%d!\n", idx);
+ return;
+ }
+
+ /* Write 64 bits and like L3C, top 16 bits are WI */
+ writeq(val, hha_pmu->base + hisi_hha_pmu_get_counter_offset(idx));
+}
+
+static void hisi_hha_pmu_write_evtype(struct hisi_pmu *hha_pmu, int idx,
+ u32 type)
+{
+ u32 reg, reg_idx, shift, val;
+
+ /*
+ * Select the appropriate event select register(HHA_EVENT_TYPEx).
+ * There are 4 event select registers for the 16 hardware counters.
+ * Event code is 8-bits and for the first 4 hardware counters,
+ * HHA_EVENT_TYPE0 is chosen. For the next 4 hardware counters,
+ * HHA_EVENT_TYPE1 is chosen and so on.
+ */
+ reg = HHA_EVENT_TYPE0 + 4 * (idx / 4);
+ reg_idx = idx % 4;
+ shift = 8 * reg_idx;
+
+ /* Write event code to HHA_EVENT_TYPEx register */
+ val = readl(hha_pmu->base + reg);
+ val &= ~(HHA_EVTYPE_NONE << shift);
+ val |= (type << shift);
+ writel(val, hha_pmu->base + reg);
+}
+
+static void hisi_hha_pmu_start_counters(struct hisi_pmu *hha_pmu)
+{
+ u32 val;
+
+ /*
+ * Set perf_enable bit in HHA_PERF_CTRL to start event
+ * counting for all enabled counters.
+ */
+ val = readl(hha_pmu->base + HHA_PERF_CTRL);
+ val |= HHA_PERF_CTRL_EN;
+ writel(val, hha_pmu->base + HHA_PERF_CTRL);
+}
+
+static void hisi_hha_pmu_stop_counters(struct hisi_pmu *hha_pmu)
+{
+ u32 val;
+
+ /*
+ * Clear perf_enable bit in HHA_PERF_CTRL to stop event
+ * counting for all enabled counters.
+ */
+ val = readl(hha_pmu->base + HHA_PERF_CTRL);
+ val &= ~(HHA_PERF_CTRL_EN);
+ writel(val, hha_pmu->base + HHA_PERF_CTRL);
+}
+
+static void hisi_hha_pmu_enable_counter(struct hisi_pmu *hha_pmu,
+ struct hw_perf_event *hwc)
+{
+ u32 val;
+
+ /* Enable counter index in HHA_EVENT_CTRL register */
+ val = readl(hha_pmu->base + HHA_EVENT_CTRL);
+ val |= (1 << hwc->idx);
+ writel(val, hha_pmu->base + HHA_EVENT_CTRL);
+}
+
+static void hisi_hha_pmu_disable_counter(struct hisi_pmu *hha_pmu,
+ struct hw_perf_event *hwc)
+{
+ u32 val;
+
+ /* Clear counter index in HHA_EVENT_CTRL register */
+ val = readl(hha_pmu->base + HHA_EVENT_CTRL);
+ val &= ~(1 << hwc->idx);
+ writel(val, hha_pmu->base + HHA_EVENT_CTRL);
+}
+
+static void hisi_hha_pmu_enable_counter_int(struct hisi_pmu *hha_pmu,
+ struct hw_perf_event *hwc)
+{
+ u32 val;
+
+ /* Write 0 to enable interrupt */
+ val = readl(hha_pmu->base + HHA_INT_MASK);
+ val &= ~(1 << hwc->idx);
+ writel(val, hha_pmu->base + HHA_INT_MASK);
+}
+
+static void hisi_hha_pmu_disable_counter_int(struct hisi_pmu *hha_pmu,
+ struct hw_perf_event *hwc)
+{
+ u32 val;
+
+ /* Write 1 to mask interrupt */
+ val = readl(hha_pmu->base + HHA_INT_MASK);
+ val |= (1 << hwc->idx);
+ writel(val, hha_pmu->base + HHA_INT_MASK);
+}
+
+static irqreturn_t hisi_hha_pmu_isr(int irq, void *dev_id)
+{
+ struct hisi_pmu *hha_pmu = dev_id;
+ struct perf_event *event;
+ unsigned long overflown;
+ int idx;
+
+ /* Read HHA_INT_STATUS register */
+ overflown = readl(hha_pmu->base + HHA_INT_STATUS);
+ if (!overflown)
+ return IRQ_NONE;
+
+ /*
+ * Find the counter index which overflowed if the bit was set
+ * and handle it
+ */
+ for_each_set_bit(idx, &overflown, HHA_NR_COUNTERS) {
+ /* Write 1 to clear the IRQ status flag */
+ writel((1 << idx), hha_pmu->base + HHA_INT_CLEAR);
+
+ /* Get the corresponding event struct */
+ event = hha_pmu->pmu_events.hw_events[idx];
+ if (!event)
+ continue;
+
+ hisi_uncore_pmu_event_update(event);
+ hisi_uncore_pmu_set_event_period(event);
+ }
+
+ return IRQ_HANDLED;
+}
+
+static int hisi_hha_pmu_init_irq(struct hisi_pmu *hha_pmu,
+ struct platform_device *pdev)
+{
+ int irq, ret;
+
+ /* Read and init IRQ */
+ irq = platform_get_irq(pdev, 0);
+ if (irq < 0) {
+ dev_err(&pdev->dev, "HHA PMU get irq fail; irq:%d\n", irq);
+ return irq;
+ }
+
+ ret = devm_request_irq(&pdev->dev, irq, hisi_hha_pmu_isr,
+ IRQF_NOBALANCING | IRQF_NO_THREAD,
+ dev_name(&pdev->dev), hha_pmu);
+ if (ret < 0) {
+ dev_err(&pdev->dev,
+ "Fail to request IRQ:%d ret:%d\n", irq, ret);
+ return ret;
+ }
+
+ hha_pmu->irq = irq;
+
+ return 0;
+}
+
+static const struct acpi_device_id hisi_hha_pmu_acpi_match[] = {
+ { "HISI0243", },
+ {},
+};
+MODULE_DEVICE_TABLE(acpi, hisi_hha_pmu_acpi_match);
+
+static int hisi_hha_pmu_init_data(struct platform_device *pdev,
+ struct hisi_pmu *hha_pmu)
+{
+ unsigned long long id;
+ struct resource *res;
+ acpi_status status;
+
+ status = acpi_evaluate_integer(ACPI_HANDLE(&pdev->dev),
+ "_UID", NULL, &id);
+ if (ACPI_FAILURE(status))
+ return -EINVAL;
+
+ hha_pmu->index_id = id;
+
+ /*
+ * Use SCCL_ID and UID to identify the HHA PMU, while
+ * SCCL_ID is in MPIDR[aff2].
+ */
+ if (device_property_read_u32(&pdev->dev, "hisilicon,scl-id",
+ &hha_pmu->sccl_id)) {
+ dev_err(&pdev->dev, "Can not read hha sccl-id!\n");
+ return -EINVAL;
+ }
+ /* HHA PMUs only share the same SCCL */
+ hha_pmu->ccl_id = -1;
+
+ res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ hha_pmu->base = devm_ioremap_resource(&pdev->dev, res);
+ if (IS_ERR(hha_pmu->base)) {
+ dev_err(&pdev->dev, "ioremap failed for hha_pmu resource\n");
+ return PTR_ERR(hha_pmu->base);
+ }
+
+ return 0;
+}
+
+static struct attribute *hisi_hha_pmu_format_attr[] = {
+ HISI_PMU_FORMAT_ATTR(event, "config:0-7"),
+ NULL,
+};
+
+static const struct attribute_group hisi_hha_pmu_format_group = {
+ .name = "format",
+ .attrs = hisi_hha_pmu_format_attr,
+};
+
+static struct attribute *hisi_hha_pmu_events_attr[] = {
+ HISI_PMU_EVENT_ATTR(rx_ops_num, 0x00),
+ HISI_PMU_EVENT_ATTR(rx_outer, 0x01),
+ HISI_PMU_EVENT_ATTR(rx_sccl, 0x02),
+ HISI_PMU_EVENT_ATTR(rx_ccix, 0x03),
+ HISI_PMU_EVENT_ATTR(rx_wbi, 0x04),
+ HISI_PMU_EVENT_ATTR(rx_wbip, 0x05),
+ HISI_PMU_EVENT_ATTR(rx_wtistash, 0x11),
+ HISI_PMU_EVENT_ATTR(rd_ddr_64b, 0x1c),
+ HISI_PMU_EVENT_ATTR(wr_dr_64b, 0x1d),
+ HISI_PMU_EVENT_ATTR(rd_ddr_128b, 0x1e),
+ HISI_PMU_EVENT_ATTR(wr_ddr_128b, 0x1f),
+ HISI_PMU_EVENT_ATTR(spill_num, 0x20),
+ HISI_PMU_EVENT_ATTR(spill_success, 0x21),
+ HISI_PMU_EVENT_ATTR(bi_num, 0x23),
+ HISI_PMU_EVENT_ATTR(mediated_num, 0x32),
+ HISI_PMU_EVENT_ATTR(tx_snp_num, 0x33),
+ HISI_PMU_EVENT_ATTR(tx_snp_outer, 0x34),
+ HISI_PMU_EVENT_ATTR(tx_snp_ccix, 0x35),
+ HISI_PMU_EVENT_ATTR(rx_snprspdata, 0x38),
+ HISI_PMU_EVENT_ATTR(rx_snprsp_outer, 0x3c),
+ HISI_PMU_EVENT_ATTR(sdir-lookup, 0x40),
+ HISI_PMU_EVENT_ATTR(edir-lookup, 0x41),
+ HISI_PMU_EVENT_ATTR(sdir-hit, 0x42),
+ HISI_PMU_EVENT_ATTR(edir-hit, 0x43),
+ HISI_PMU_EVENT_ATTR(sdir-home-migrate, 0x4c),
+ HISI_PMU_EVENT_ATTR(edir-home-migrate, 0x4d),
+ NULL,
+};
+
+static const struct attribute_group hisi_hha_pmu_events_group = {
+ .name = "events",
+ .attrs = hisi_hha_pmu_events_attr,
+};
+
+static DEVICE_ATTR(cpumask, 0444, hisi_cpumask_sysfs_show, NULL);
+
+static struct attribute *hisi_hha_pmu_cpumask_attrs[] = {
+ &dev_attr_cpumask.attr,
+ NULL,
+};
+
+static const struct attribute_group hisi_hha_pmu_cpumask_attr_group = {
+ .attrs = hisi_hha_pmu_cpumask_attrs,
+};
+
+static const struct attribute_group *hisi_hha_pmu_attr_groups[] = {
+ &hisi_hha_pmu_format_group,
+ &hisi_hha_pmu_events_group,
+ &hisi_hha_pmu_cpumask_attr_group,
+ NULL,
+};
+
+static const struct hisi_uncore_ops hisi_uncore_hha_ops = {
+ .write_evtype = hisi_hha_pmu_write_evtype,
+ .get_event_idx = hisi_uncore_pmu_get_event_idx,
+ .start_counters = hisi_hha_pmu_start_counters,
+ .stop_counters = hisi_hha_pmu_stop_counters,
+ .enable_counter = hisi_hha_pmu_enable_counter,
+ .disable_counter = hisi_hha_pmu_disable_counter,
+ .enable_counter_int = hisi_hha_pmu_enable_counter_int,
+ .disable_counter_int = hisi_hha_pmu_disable_counter_int,
+ .write_counter = hisi_hha_pmu_write_counter,
+ .read_counter = hisi_hha_pmu_read_counter,
+};
+
+static int hisi_hha_pmu_dev_probe(struct platform_device *pdev,
+ struct hisi_pmu *hha_pmu)
+{
+ int ret;
+
+ ret = hisi_hha_pmu_init_data(pdev, hha_pmu);
+ if (ret)
+ return ret;
+
+ ret = hisi_hha_pmu_init_irq(hha_pmu, pdev);
+ if (ret)
+ return ret;
+
+ hha_pmu->num_counters = HHA_NR_COUNTERS;
+ hha_pmu->counter_bits = 48;
+ hha_pmu->ops = &hisi_uncore_hha_ops;
+ hha_pmu->dev = &pdev->dev;
+ hha_pmu->on_cpu = -1;
+ hha_pmu->check_event = 0x65;
+
+ return 0;
+}
+
+static int hisi_hha_pmu_probe(struct platform_device *pdev)
+{
+ struct hisi_pmu *hha_pmu;
+ char *name;
+ int ret;
+
+ hha_pmu = devm_kzalloc(&pdev->dev, sizeof(*hha_pmu), GFP_KERNEL);
+ if (!hha_pmu)
+ return -ENOMEM;
+
+ platform_set_drvdata(pdev, hha_pmu);
+
+ ret = hisi_hha_pmu_dev_probe(pdev, hha_pmu);
+ if (ret)
+ return ret;
+
+ ret = cpuhp_state_add_instance(CPUHP_AP_PERF_ARM_HISI_HHA_ONLINE,
+ &hha_pmu->node);
+ if (ret) {
+ dev_err(&pdev->dev, "Error %d registering hotplug\n", ret);
+ return ret;
+ }
+
+ name = devm_kasprintf(&pdev->dev, GFP_KERNEL, "hisi_sccl%u_hha%u",
+ hha_pmu->sccl_id, hha_pmu->index_id);
+ hha_pmu->pmu = (struct pmu) {
+ .name = name,
+ .task_ctx_nr = perf_invalid_context,
+ .event_init = hisi_uncore_pmu_event_init,
+ .pmu_enable = hisi_uncore_pmu_enable,
+ .pmu_disable = hisi_uncore_pmu_disable,
+ .add = hisi_uncore_pmu_add,
+ .del = hisi_uncore_pmu_del,
+ .start = hisi_uncore_pmu_start,
+ .stop = hisi_uncore_pmu_stop,
+ .read = hisi_uncore_pmu_read,
+ .attr_groups = hisi_hha_pmu_attr_groups,
+ };
+
+ ret = perf_pmu_register(&hha_pmu->pmu, name, -1);
+ if (ret) {
+ dev_err(hha_pmu->dev, "HHA PMU register failed!\n");
+ cpuhp_state_remove_instance(CPUHP_AP_PERF_ARM_HISI_HHA_ONLINE,
+ &hha_pmu->node);
+ }
+
+ return ret;
+}
+
+static int hisi_hha_pmu_remove(struct platform_device *pdev)
+{
+ struct hisi_pmu *hha_pmu = platform_get_drvdata(pdev);
+
+ perf_pmu_unregister(&hha_pmu->pmu);
+ cpuhp_state_remove_instance(CPUHP_AP_PERF_ARM_HISI_HHA_ONLINE,
+ &hha_pmu->node);
+
+ return 0;
+}
+
+static struct platform_driver hisi_hha_pmu_driver = {
+ .driver = {
+ .name = "hisi_hha_pmu",
+ .acpi_match_table = ACPI_PTR(hisi_hha_pmu_acpi_match),
+ },
+ .probe = hisi_hha_pmu_probe,
+ .remove = hisi_hha_pmu_remove,
+};
+
+static int __init hisi_hha_pmu_module_init(void)
+{
+ int ret;
+
+ ret = cpuhp_setup_state_multi(CPUHP_AP_PERF_ARM_HISI_HHA_ONLINE,
+ "AP_PERF_ARM_HISI_HHA_ONLINE",
+ hisi_uncore_pmu_online_cpu,
+ hisi_uncore_pmu_offline_cpu);
+ if (ret) {
+ pr_err("HHA PMU: Error setup hotplug, ret = %d;\n", ret);
+ return ret;
+ }
+
+ ret = platform_driver_register(&hisi_hha_pmu_driver);
+ if (ret)
+ cpuhp_remove_multi_state(CPUHP_AP_PERF_ARM_HISI_HHA_ONLINE);
+
+ return ret;
+}
+module_init(hisi_hha_pmu_module_init);
+
+static void __exit hisi_hha_pmu_module_exit(void)
+{
+ platform_driver_unregister(&hisi_hha_pmu_driver);
+ cpuhp_remove_multi_state(CPUHP_AP_PERF_ARM_HISI_HHA_ONLINE);
+}
+module_exit(hisi_hha_pmu_module_exit);
+
+MODULE_DESCRIPTION("HiSilicon SoC HHA uncore PMU driver");
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("Shaokun Zhang <zhangshaokun@hisilicon.com>");
+MODULE_AUTHOR("Anurup M <anurup.m@huawei.com>");
diff --git a/drivers/perf/hisilicon/hisi_uncore_l3c_pmu.c b/drivers/perf/hisilicon/hisi_uncore_l3c_pmu.c
new file mode 100644
index 000000000000..0bde5d919b2e
--- /dev/null
+++ b/drivers/perf/hisilicon/hisi_uncore_l3c_pmu.c
@@ -0,0 +1,463 @@
+/*
+ * HiSilicon SoC L3C uncore Hardware event counters support
+ *
+ * Copyright (C) 2017 Hisilicon Limited
+ * Author: Anurup M <anurup.m@huawei.com>
+ * Shaokun Zhang <zhangshaokun@hisilicon.com>
+ *
+ * This code is based on the uncore PMUs like arm-cci and arm-ccn.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/acpi.h>
+#include <linux/bug.h>
+#include <linux/cpuhotplug.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/list.h>
+#include <linux/platform_device.h>
+#include <linux/smp.h>
+
+#include "hisi_uncore_pmu.h"
+
+/* L3C register definition */
+#define L3C_PERF_CTRL 0x0408
+#define L3C_INT_MASK 0x0800
+#define L3C_INT_STATUS 0x0808
+#define L3C_INT_CLEAR 0x080c
+#define L3C_EVENT_CTRL 0x1c00
+#define L3C_EVENT_TYPE0 0x1d00
+/*
+ * Each counter is 48-bits and [48:63] are reserved
+ * which are Read-As-Zero and Writes-Ignored.
+ */
+#define L3C_CNTR0_LOWER 0x1e00
+
+/* L3C has 8-counters */
+#define L3C_NR_COUNTERS 0x8
+
+#define L3C_PERF_CTRL_EN 0x20000
+#define L3C_EVTYPE_NONE 0xff
+
+/*
+ * Select the counter register offset using the counter index
+ */
+static u32 hisi_l3c_pmu_get_counter_offset(int cntr_idx)
+{
+ return (L3C_CNTR0_LOWER + (cntr_idx * 8));
+}
+
+static u64 hisi_l3c_pmu_read_counter(struct hisi_pmu *l3c_pmu,
+ struct hw_perf_event *hwc)
+{
+ u32 idx = hwc->idx;
+
+ if (!hisi_uncore_pmu_counter_valid(l3c_pmu, idx)) {
+ dev_err(l3c_pmu->dev, "Unsupported event index:%d!\n", idx);
+ return 0;
+ }
+
+ /* Read 64-bits and the upper 16 bits are RAZ */
+ return readq(l3c_pmu->base + hisi_l3c_pmu_get_counter_offset(idx));
+}
+
+static void hisi_l3c_pmu_write_counter(struct hisi_pmu *l3c_pmu,
+ struct hw_perf_event *hwc, u64 val)
+{
+ u32 idx = hwc->idx;
+
+ if (!hisi_uncore_pmu_counter_valid(l3c_pmu, idx)) {
+ dev_err(l3c_pmu->dev, "Unsupported event index:%d!\n", idx);
+ return;
+ }
+
+ /* Write 64-bits and the upper 16 bits are WI */
+ writeq(val, l3c_pmu->base + hisi_l3c_pmu_get_counter_offset(idx));
+}
+
+static void hisi_l3c_pmu_write_evtype(struct hisi_pmu *l3c_pmu, int idx,
+ u32 type)
+{
+ u32 reg, reg_idx, shift, val;
+
+ /*
+ * Select the appropriate event select register(L3C_EVENT_TYPE0/1).
+ * There are 2 event select registers for the 8 hardware counters.
+ * Event code is 8-bits and for the former 4 hardware counters,
+ * L3C_EVENT_TYPE0 is chosen. For the latter 4 hardware counters,
+ * L3C_EVENT_TYPE1 is chosen.
+ */
+ reg = L3C_EVENT_TYPE0 + (idx / 4) * 4;
+ reg_idx = idx % 4;
+ shift = 8 * reg_idx;
+
+ /* Write event code to L3C_EVENT_TYPEx Register */
+ val = readl(l3c_pmu->base + reg);
+ val &= ~(L3C_EVTYPE_NONE << shift);
+ val |= (type << shift);
+ writel(val, l3c_pmu->base + reg);
+}
+
+static void hisi_l3c_pmu_start_counters(struct hisi_pmu *l3c_pmu)
+{
+ u32 val;
+
+ /*
+ * Set perf_enable bit in L3C_PERF_CTRL register to start counting
+ * for all enabled counters.
+ */
+ val = readl(l3c_pmu->base + L3C_PERF_CTRL);
+ val |= L3C_PERF_CTRL_EN;
+ writel(val, l3c_pmu->base + L3C_PERF_CTRL);
+}
+
+static void hisi_l3c_pmu_stop_counters(struct hisi_pmu *l3c_pmu)
+{
+ u32 val;
+
+ /*
+ * Clear perf_enable bit in L3C_PERF_CTRL register to stop counting
+ * for all enabled counters.
+ */
+ val = readl(l3c_pmu->base + L3C_PERF_CTRL);
+ val &= ~(L3C_PERF_CTRL_EN);
+ writel(val, l3c_pmu->base + L3C_PERF_CTRL);
+}
+
+static void hisi_l3c_pmu_enable_counter(struct hisi_pmu *l3c_pmu,
+ struct hw_perf_event *hwc)
+{
+ u32 val;
+
+ /* Enable counter index in L3C_EVENT_CTRL register */
+ val = readl(l3c_pmu->base + L3C_EVENT_CTRL);
+ val |= (1 << hwc->idx);
+ writel(val, l3c_pmu->base + L3C_EVENT_CTRL);
+}
+
+static void hisi_l3c_pmu_disable_counter(struct hisi_pmu *l3c_pmu,
+ struct hw_perf_event *hwc)
+{
+ u32 val;
+
+ /* Clear counter index in L3C_EVENT_CTRL register */
+ val = readl(l3c_pmu->base + L3C_EVENT_CTRL);
+ val &= ~(1 << hwc->idx);
+ writel(val, l3c_pmu->base + L3C_EVENT_CTRL);
+}
+
+static void hisi_l3c_pmu_enable_counter_int(struct hisi_pmu *l3c_pmu,
+ struct hw_perf_event *hwc)
+{
+ u32 val;
+
+ val = readl(l3c_pmu->base + L3C_INT_MASK);
+ /* Write 0 to enable interrupt */
+ val &= ~(1 << hwc->idx);
+ writel(val, l3c_pmu->base + L3C_INT_MASK);
+}
+
+static void hisi_l3c_pmu_disable_counter_int(struct hisi_pmu *l3c_pmu,
+ struct hw_perf_event *hwc)
+{
+ u32 val;
+
+ val = readl(l3c_pmu->base + L3C_INT_MASK);
+ /* Write 1 to mask interrupt */
+ val |= (1 << hwc->idx);
+ writel(val, l3c_pmu->base + L3C_INT_MASK);
+}
+
+static irqreturn_t hisi_l3c_pmu_isr(int irq, void *dev_id)
+{
+ struct hisi_pmu *l3c_pmu = dev_id;
+ struct perf_event *event;
+ unsigned long overflown;
+ int idx;
+
+ /* Read L3C_INT_STATUS register */
+ overflown = readl(l3c_pmu->base + L3C_INT_STATUS);
+ if (!overflown)
+ return IRQ_NONE;
+
+ /*
+ * Find the counter index which overflowed if the bit was set
+ * and handle it.
+ */
+ for_each_set_bit(idx, &overflown, L3C_NR_COUNTERS) {
+ /* Write 1 to clear the IRQ status flag */
+ writel((1 << idx), l3c_pmu->base + L3C_INT_CLEAR);
+
+ /* Get the corresponding event struct */
+ event = l3c_pmu->pmu_events.hw_events[idx];
+ if (!event)
+ continue;
+
+ hisi_uncore_pmu_event_update(event);
+ hisi_uncore_pmu_set_event_period(event);
+ }
+
+ return IRQ_HANDLED;
+}
+
+static int hisi_l3c_pmu_init_irq(struct hisi_pmu *l3c_pmu,
+ struct platform_device *pdev)
+{
+ int irq, ret;
+
+ /* Read and init IRQ */
+ irq = platform_get_irq(pdev, 0);
+ if (irq < 0) {
+ dev_err(&pdev->dev, "L3C PMU get irq fail; irq:%d\n", irq);
+ return irq;
+ }
+
+ ret = devm_request_irq(&pdev->dev, irq, hisi_l3c_pmu_isr,
+ IRQF_NOBALANCING | IRQF_NO_THREAD,
+ dev_name(&pdev->dev), l3c_pmu);
+ if (ret < 0) {
+ dev_err(&pdev->dev,
+ "Fail to request IRQ:%d ret:%d\n", irq, ret);
+ return ret;
+ }
+
+ l3c_pmu->irq = irq;
+
+ return 0;
+}
+
+static const struct acpi_device_id hisi_l3c_pmu_acpi_match[] = {
+ { "HISI0213", },
+ {},
+};
+MODULE_DEVICE_TABLE(acpi, hisi_l3c_pmu_acpi_match);
+
+static int hisi_l3c_pmu_init_data(struct platform_device *pdev,
+ struct hisi_pmu *l3c_pmu)
+{
+ unsigned long long id;
+ struct resource *res;
+ acpi_status status;
+
+ status = acpi_evaluate_integer(ACPI_HANDLE(&pdev->dev),
+ "_UID", NULL, &id);
+ if (ACPI_FAILURE(status))
+ return -EINVAL;
+
+ l3c_pmu->index_id = id;
+
+ /*
+ * Use the SCCL_ID and CCL_ID to identify the L3C PMU, while
+ * SCCL_ID is in MPIDR[aff2] and CCL_ID is in MPIDR[aff1].
+ */
+ if (device_property_read_u32(&pdev->dev, "hisilicon,scl-id",
+ &l3c_pmu->sccl_id)) {
+ dev_err(&pdev->dev, "Can not read l3c sccl-id!\n");
+ return -EINVAL;
+ }
+
+ if (device_property_read_u32(&pdev->dev, "hisilicon,ccl-id",
+ &l3c_pmu->ccl_id)) {
+ dev_err(&pdev->dev, "Can not read l3c ccl-id!\n");
+ return -EINVAL;
+ }
+
+ res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ l3c_pmu->base = devm_ioremap_resource(&pdev->dev, res);
+ if (IS_ERR(l3c_pmu->base)) {
+ dev_err(&pdev->dev, "ioremap failed for l3c_pmu resource\n");
+ return PTR_ERR(l3c_pmu->base);
+ }
+
+ return 0;
+}
+
+static struct attribute *hisi_l3c_pmu_format_attr[] = {
+ HISI_PMU_FORMAT_ATTR(event, "config:0-7"),
+ NULL,
+};
+
+static const struct attribute_group hisi_l3c_pmu_format_group = {
+ .name = "format",
+ .attrs = hisi_l3c_pmu_format_attr,
+};
+
+static struct attribute *hisi_l3c_pmu_events_attr[] = {
+ HISI_PMU_EVENT_ATTR(rd_cpipe, 0x00),
+ HISI_PMU_EVENT_ATTR(wr_cpipe, 0x01),
+ HISI_PMU_EVENT_ATTR(rd_hit_cpipe, 0x02),
+ HISI_PMU_EVENT_ATTR(wr_hit_cpipe, 0x03),
+ HISI_PMU_EVENT_ATTR(victim_num, 0x04),
+ HISI_PMU_EVENT_ATTR(rd_spipe, 0x20),
+ HISI_PMU_EVENT_ATTR(wr_spipe, 0x21),
+ HISI_PMU_EVENT_ATTR(rd_hit_spipe, 0x22),
+ HISI_PMU_EVENT_ATTR(wr_hit_spipe, 0x23),
+ HISI_PMU_EVENT_ATTR(back_invalid, 0x29),
+ HISI_PMU_EVENT_ATTR(retry_cpu, 0x40),
+ HISI_PMU_EVENT_ATTR(retry_ring, 0x41),
+ HISI_PMU_EVENT_ATTR(prefetch_drop, 0x42),
+ NULL,
+};
+
+static const struct attribute_group hisi_l3c_pmu_events_group = {
+ .name = "events",
+ .attrs = hisi_l3c_pmu_events_attr,
+};
+
+static DEVICE_ATTR(cpumask, 0444, hisi_cpumask_sysfs_show, NULL);
+
+static struct attribute *hisi_l3c_pmu_cpumask_attrs[] = {
+ &dev_attr_cpumask.attr,
+ NULL,
+};
+
+static const struct attribute_group hisi_l3c_pmu_cpumask_attr_group = {
+ .attrs = hisi_l3c_pmu_cpumask_attrs,
+};
+
+static const struct attribute_group *hisi_l3c_pmu_attr_groups[] = {
+ &hisi_l3c_pmu_format_group,
+ &hisi_l3c_pmu_events_group,
+ &hisi_l3c_pmu_cpumask_attr_group,
+ NULL,
+};
+
+static const struct hisi_uncore_ops hisi_uncore_l3c_ops = {
+ .write_evtype = hisi_l3c_pmu_write_evtype,
+ .get_event_idx = hisi_uncore_pmu_get_event_idx,
+ .start_counters = hisi_l3c_pmu_start_counters,
+ .stop_counters = hisi_l3c_pmu_stop_counters,
+ .enable_counter = hisi_l3c_pmu_enable_counter,
+ .disable_counter = hisi_l3c_pmu_disable_counter,
+ .enable_counter_int = hisi_l3c_pmu_enable_counter_int,
+ .disable_counter_int = hisi_l3c_pmu_disable_counter_int,
+ .write_counter = hisi_l3c_pmu_write_counter,
+ .read_counter = hisi_l3c_pmu_read_counter,
+};
+
+static int hisi_l3c_pmu_dev_probe(struct platform_device *pdev,
+ struct hisi_pmu *l3c_pmu)
+{
+ int ret;
+
+ ret = hisi_l3c_pmu_init_data(pdev, l3c_pmu);
+ if (ret)
+ return ret;
+
+ ret = hisi_l3c_pmu_init_irq(l3c_pmu, pdev);
+ if (ret)
+ return ret;
+
+ l3c_pmu->num_counters = L3C_NR_COUNTERS;
+ l3c_pmu->counter_bits = 48;
+ l3c_pmu->ops = &hisi_uncore_l3c_ops;
+ l3c_pmu->dev = &pdev->dev;
+ l3c_pmu->on_cpu = -1;
+ l3c_pmu->check_event = 0x59;
+
+ return 0;
+}
+
+static int hisi_l3c_pmu_probe(struct platform_device *pdev)
+{
+ struct hisi_pmu *l3c_pmu;
+ char *name;
+ int ret;
+
+ l3c_pmu = devm_kzalloc(&pdev->dev, sizeof(*l3c_pmu), GFP_KERNEL);
+ if (!l3c_pmu)
+ return -ENOMEM;
+
+ platform_set_drvdata(pdev, l3c_pmu);
+
+ ret = hisi_l3c_pmu_dev_probe(pdev, l3c_pmu);
+ if (ret)
+ return ret;
+
+ ret = cpuhp_state_add_instance(CPUHP_AP_PERF_ARM_HISI_L3_ONLINE,
+ &l3c_pmu->node);
+ if (ret) {
+ dev_err(&pdev->dev, "Error %d registering hotplug\n", ret);
+ return ret;
+ }
+
+ name = devm_kasprintf(&pdev->dev, GFP_KERNEL, "hisi_sccl%u_l3c%u",
+ l3c_pmu->sccl_id, l3c_pmu->index_id);
+ l3c_pmu->pmu = (struct pmu) {
+ .name = name,
+ .task_ctx_nr = perf_invalid_context,
+ .event_init = hisi_uncore_pmu_event_init,
+ .pmu_enable = hisi_uncore_pmu_enable,
+ .pmu_disable = hisi_uncore_pmu_disable,
+ .add = hisi_uncore_pmu_add,
+ .del = hisi_uncore_pmu_del,
+ .start = hisi_uncore_pmu_start,
+ .stop = hisi_uncore_pmu_stop,
+ .read = hisi_uncore_pmu_read,
+ .attr_groups = hisi_l3c_pmu_attr_groups,
+ };
+
+ ret = perf_pmu_register(&l3c_pmu->pmu, name, -1);
+ if (ret) {
+ dev_err(l3c_pmu->dev, "L3C PMU register failed!\n");
+ cpuhp_state_remove_instance(CPUHP_AP_PERF_ARM_HISI_L3_ONLINE,
+ &l3c_pmu->node);
+ }
+
+ return ret;
+}
+
+static int hisi_l3c_pmu_remove(struct platform_device *pdev)
+{
+ struct hisi_pmu *l3c_pmu = platform_get_drvdata(pdev);
+
+ perf_pmu_unregister(&l3c_pmu->pmu);
+ cpuhp_state_remove_instance(CPUHP_AP_PERF_ARM_HISI_L3_ONLINE,
+ &l3c_pmu->node);
+
+ return 0;
+}
+
+static struct platform_driver hisi_l3c_pmu_driver = {
+ .driver = {
+ .name = "hisi_l3c_pmu",
+ .acpi_match_table = ACPI_PTR(hisi_l3c_pmu_acpi_match),
+ },
+ .probe = hisi_l3c_pmu_probe,
+ .remove = hisi_l3c_pmu_remove,
+};
+
+static int __init hisi_l3c_pmu_module_init(void)
+{
+ int ret;
+
+ ret = cpuhp_setup_state_multi(CPUHP_AP_PERF_ARM_HISI_L3_ONLINE,
+ "AP_PERF_ARM_HISI_L3_ONLINE",
+ hisi_uncore_pmu_online_cpu,
+ hisi_uncore_pmu_offline_cpu);
+ if (ret) {
+ pr_err("L3C PMU: Error setup hotplug, ret = %d\n", ret);
+ return ret;
+ }
+
+ ret = platform_driver_register(&hisi_l3c_pmu_driver);
+ if (ret)
+ cpuhp_remove_multi_state(CPUHP_AP_PERF_ARM_HISI_L3_ONLINE);
+
+ return ret;
+}
+module_init(hisi_l3c_pmu_module_init);
+
+static void __exit hisi_l3c_pmu_module_exit(void)
+{
+ platform_driver_unregister(&hisi_l3c_pmu_driver);
+ cpuhp_remove_multi_state(CPUHP_AP_PERF_ARM_HISI_L3_ONLINE);
+}
+module_exit(hisi_l3c_pmu_module_exit);
+
+MODULE_DESCRIPTION("HiSilicon SoC L3C uncore PMU driver");
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("Anurup M <anurup.m@huawei.com>");
+MODULE_AUTHOR("Shaokun Zhang <zhangshaokun@hisilicon.com>");
diff --git a/drivers/perf/hisilicon/hisi_uncore_pmu.c b/drivers/perf/hisilicon/hisi_uncore_pmu.c
new file mode 100644
index 000000000000..7ed24b954422
--- /dev/null
+++ b/drivers/perf/hisilicon/hisi_uncore_pmu.c
@@ -0,0 +1,447 @@
+/*
+ * HiSilicon SoC Hardware event counters support
+ *
+ * Copyright (C) 2017 Hisilicon Limited
+ * Author: Anurup M <anurup.m@huawei.com>
+ * Shaokun Zhang <zhangshaokun@hisilicon.com>
+ *
+ * This code is based on the uncore PMUs like arm-cci and arm-ccn.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/bitmap.h>
+#include <linux/bitops.h>
+#include <linux/bug.h>
+#include <linux/err.h>
+#include <linux/errno.h>
+#include <linux/interrupt.h>
+
+#include <asm/local64.h>
+
+#include "hisi_uncore_pmu.h"
+
+#define HISI_GET_EVENTID(ev) (ev->hw.config_base & 0xff)
+#define HISI_MAX_PERIOD(nr) (BIT_ULL(nr) - 1)
+
+/*
+ * PMU format attributes
+ */
+ssize_t hisi_format_sysfs_show(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct dev_ext_attribute *eattr;
+
+ eattr = container_of(attr, struct dev_ext_attribute, attr);
+
+ return sprintf(buf, "%s\n", (char *)eattr->var);
+}
+
+/*
+ * PMU event attributes
+ */
+ssize_t hisi_event_sysfs_show(struct device *dev,
+ struct device_attribute *attr, char *page)
+{
+ struct dev_ext_attribute *eattr;
+
+ eattr = container_of(attr, struct dev_ext_attribute, attr);
+
+ return sprintf(page, "config=0x%lx\n", (unsigned long)eattr->var);
+}
+
+/*
+ * sysfs cpumask attributes. For uncore PMU, we only have a single CPU to show
+ */
+ssize_t hisi_cpumask_sysfs_show(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct hisi_pmu *hisi_pmu = to_hisi_pmu(dev_get_drvdata(dev));
+
+ return sprintf(buf, "%d\n", hisi_pmu->on_cpu);
+}
+
+static bool hisi_validate_event_group(struct perf_event *event)
+{
+ struct perf_event *sibling, *leader = event->group_leader;
+ struct hisi_pmu *hisi_pmu = to_hisi_pmu(event->pmu);
+ /* Include count for the event */
+ int counters = 1;
+
+ if (!is_software_event(leader)) {
+ /*
+ * We must NOT create groups containing mixed PMUs, although
+ * software events are acceptable
+ */
+ if (leader->pmu != event->pmu)
+ return false;
+
+ /* Increment counter for the leader */
+ if (leader != event)
+ counters++;
+ }
+
+ list_for_each_entry(sibling, &event->group_leader->sibling_list,
+ group_entry) {
+ if (is_software_event(sibling))
+ continue;
+ if (sibling->pmu != event->pmu)
+ return false;
+ /* Increment counter for each sibling */
+ counters++;
+ }
+
+ /* The group can not count events more than the counters in the HW */
+ return counters <= hisi_pmu->num_counters;
+}
+
+int hisi_uncore_pmu_counter_valid(struct hisi_pmu *hisi_pmu, int idx)
+{
+ return idx >= 0 && idx < hisi_pmu->num_counters;
+}
+
+int hisi_uncore_pmu_get_event_idx(struct perf_event *event)
+{
+ struct hisi_pmu *hisi_pmu = to_hisi_pmu(event->pmu);
+ unsigned long *used_mask = hisi_pmu->pmu_events.used_mask;
+ u32 num_counters = hisi_pmu->num_counters;
+ int idx;
+
+ idx = find_first_zero_bit(used_mask, num_counters);
+ if (idx == num_counters)
+ return -EAGAIN;
+
+ set_bit(idx, used_mask);
+
+ return idx;
+}
+
+static void hisi_uncore_pmu_clear_event_idx(struct hisi_pmu *hisi_pmu, int idx)
+{
+ if (!hisi_uncore_pmu_counter_valid(hisi_pmu, idx)) {
+ dev_err(hisi_pmu->dev, "Unsupported event index:%d!\n", idx);
+ return;
+ }
+
+ clear_bit(idx, hisi_pmu->pmu_events.used_mask);
+}
+
+int hisi_uncore_pmu_event_init(struct perf_event *event)
+{
+ struct hw_perf_event *hwc = &event->hw;
+ struct hisi_pmu *hisi_pmu;
+
+ if (event->attr.type != event->pmu->type)
+ return -ENOENT;
+
+ /*
+ * We do not support sampling as the counters are all
+ * shared by all CPU cores in a CPU die(SCCL). Also we
+ * do not support attach to a task(per-process mode)
+ */
+ if (is_sampling_event(event) || event->attach_state & PERF_ATTACH_TASK)
+ return -EOPNOTSUPP;
+
+ /* counters do not have these bits */
+ if (event->attr.exclude_user ||
+ event->attr.exclude_kernel ||
+ event->attr.exclude_host ||
+ event->attr.exclude_guest ||
+ event->attr.exclude_hv ||
+ event->attr.exclude_idle)
+ return -EINVAL;
+
+ /*
+ * The uncore counters not specific to any CPU, so cannot
+ * support per-task
+ */
+ if (event->cpu < 0)
+ return -EINVAL;
+
+ /*
+ * Validate if the events in group does not exceed the
+ * available counters in hardware.
+ */
+ if (!hisi_validate_event_group(event))
+ return -EINVAL;
+
+ hisi_pmu = to_hisi_pmu(event->pmu);
+ if (event->attr.config > hisi_pmu->check_event)
+ return -EINVAL;
+
+ if (hisi_pmu->on_cpu == -1)
+ return -EINVAL;
+ /*
+ * We don't assign an index until we actually place the event onto
+ * hardware. Use -1 to signify that we haven't decided where to put it
+ * yet.
+ */
+ hwc->idx = -1;
+ hwc->config_base = event->attr.config;
+
+ /* Enforce to use the same CPU for all events in this PMU */
+ event->cpu = hisi_pmu->on_cpu;
+
+ return 0;
+}
+
+/*
+ * Set the counter to count the event that we're interested in,
+ * and enable interrupt and counter.
+ */
+static void hisi_uncore_pmu_enable_event(struct perf_event *event)
+{
+ struct hisi_pmu *hisi_pmu = to_hisi_pmu(event->pmu);
+ struct hw_perf_event *hwc = &event->hw;
+
+ hisi_pmu->ops->write_evtype(hisi_pmu, hwc->idx,
+ HISI_GET_EVENTID(event));
+
+ hisi_pmu->ops->enable_counter_int(hisi_pmu, hwc);
+ hisi_pmu->ops->enable_counter(hisi_pmu, hwc);
+}
+
+/*
+ * Disable counter and interrupt.
+ */
+static void hisi_uncore_pmu_disable_event(struct perf_event *event)
+{
+ struct hisi_pmu *hisi_pmu = to_hisi_pmu(event->pmu);
+ struct hw_perf_event *hwc = &event->hw;
+
+ hisi_pmu->ops->disable_counter(hisi_pmu, hwc);
+ hisi_pmu->ops->disable_counter_int(hisi_pmu, hwc);
+}
+
+void hisi_uncore_pmu_set_event_period(struct perf_event *event)
+{
+ struct hisi_pmu *hisi_pmu = to_hisi_pmu(event->pmu);
+ struct hw_perf_event *hwc = &event->hw;
+
+ /*
+ * The HiSilicon PMU counters support 32 bits or 48 bits, depending on
+ * the PMU. We reduce it to 2^(counter_bits - 1) to account for the
+ * extreme interrupt latency. So we could hopefully handle the overflow
+ * interrupt before another 2^(counter_bits - 1) events occur and the
+ * counter overtakes its previous value.
+ */
+ u64 val = BIT_ULL(hisi_pmu->counter_bits - 1);
+
+ local64_set(&hwc->prev_count, val);
+ /* Write start value to the hardware event counter */
+ hisi_pmu->ops->write_counter(hisi_pmu, hwc, val);
+}
+
+void hisi_uncore_pmu_event_update(struct perf_event *event)
+{
+ struct hisi_pmu *hisi_pmu = to_hisi_pmu(event->pmu);
+ struct hw_perf_event *hwc = &event->hw;
+ u64 delta, prev_raw_count, new_raw_count;
+
+ do {
+ /* Read the count from the counter register */
+ new_raw_count = hisi_pmu->ops->read_counter(hisi_pmu, hwc);
+ prev_raw_count = local64_read(&hwc->prev_count);
+ } while (local64_cmpxchg(&hwc->prev_count, prev_raw_count,
+ new_raw_count) != prev_raw_count);
+ /*
+ * compute the delta
+ */
+ delta = (new_raw_count - prev_raw_count) &
+ HISI_MAX_PERIOD(hisi_pmu->counter_bits);
+ local64_add(delta, &event->count);
+}
+
+void hisi_uncore_pmu_start(struct perf_event *event, int flags)
+{
+ struct hisi_pmu *hisi_pmu = to_hisi_pmu(event->pmu);
+ struct hw_perf_event *hwc = &event->hw;
+
+ if (WARN_ON_ONCE(!(hwc->state & PERF_HES_STOPPED)))
+ return;
+
+ WARN_ON_ONCE(!(hwc->state & PERF_HES_UPTODATE));
+ hwc->state = 0;
+ hisi_uncore_pmu_set_event_period(event);
+
+ if (flags & PERF_EF_RELOAD) {
+ u64 prev_raw_count = local64_read(&hwc->prev_count);
+
+ hisi_pmu->ops->write_counter(hisi_pmu, hwc, prev_raw_count);
+ }
+
+ hisi_uncore_pmu_enable_event(event);
+ perf_event_update_userpage(event);
+}
+
+void hisi_uncore_pmu_stop(struct perf_event *event, int flags)
+{
+ struct hw_perf_event *hwc = &event->hw;
+
+ hisi_uncore_pmu_disable_event(event);
+ WARN_ON_ONCE(hwc->state & PERF_HES_STOPPED);
+ hwc->state |= PERF_HES_STOPPED;
+
+ if (hwc->state & PERF_HES_UPTODATE)
+ return;
+
+ /* Read hardware counter and update the perf counter statistics */
+ hisi_uncore_pmu_event_update(event);
+ hwc->state |= PERF_HES_UPTODATE;
+}
+
+int hisi_uncore_pmu_add(struct perf_event *event, int flags)
+{
+ struct hisi_pmu *hisi_pmu = to_hisi_pmu(event->pmu);
+ struct hw_perf_event *hwc = &event->hw;
+ int idx;
+
+ hwc->state = PERF_HES_STOPPED | PERF_HES_UPTODATE;
+
+ /* Get an available counter index for counting */
+ idx = hisi_pmu->ops->get_event_idx(event);
+ if (idx < 0)
+ return idx;
+
+ event->hw.idx = idx;
+ hisi_pmu->pmu_events.hw_events[idx] = event;
+
+ if (flags & PERF_EF_START)
+ hisi_uncore_pmu_start(event, PERF_EF_RELOAD);
+
+ return 0;
+}
+
+void hisi_uncore_pmu_del(struct perf_event *event, int flags)
+{
+ struct hisi_pmu *hisi_pmu = to_hisi_pmu(event->pmu);
+ struct hw_perf_event *hwc = &event->hw;
+
+ hisi_uncore_pmu_stop(event, PERF_EF_UPDATE);
+ hisi_uncore_pmu_clear_event_idx(hisi_pmu, hwc->idx);
+ perf_event_update_userpage(event);
+ hisi_pmu->pmu_events.hw_events[hwc->idx] = NULL;
+}
+
+void hisi_uncore_pmu_read(struct perf_event *event)
+{
+ /* Read hardware counter and update the perf counter statistics */
+ hisi_uncore_pmu_event_update(event);
+}
+
+void hisi_uncore_pmu_enable(struct pmu *pmu)
+{
+ struct hisi_pmu *hisi_pmu = to_hisi_pmu(pmu);
+ int enabled = bitmap_weight(hisi_pmu->pmu_events.used_mask,
+ hisi_pmu->num_counters);
+
+ if (!enabled)
+ return;
+
+ hisi_pmu->ops->start_counters(hisi_pmu);
+}
+
+void hisi_uncore_pmu_disable(struct pmu *pmu)
+{
+ struct hisi_pmu *hisi_pmu = to_hisi_pmu(pmu);
+
+ hisi_pmu->ops->stop_counters(hisi_pmu);
+}
+
+/*
+ * Read Super CPU cluster and CPU cluster ID from MPIDR_EL1.
+ * If multi-threading is supported, SCCL_ID is in MPIDR[aff3] and CCL_ID
+ * is in MPIDR[aff2]; if not, SCCL_ID is in MPIDR[aff2] and CCL_ID is
+ * in MPIDR[aff1]. If this changes in future, this shall be updated.
+ */
+static void hisi_read_sccl_and_ccl_id(int *sccl_id, int *ccl_id)
+{
+ u64 mpidr = read_cpuid_mpidr();
+
+ if (mpidr & MPIDR_MT_BITMASK) {
+ if (sccl_id)
+ *sccl_id = MPIDR_AFFINITY_LEVEL(mpidr, 3);
+ if (ccl_id)
+ *ccl_id = MPIDR_AFFINITY_LEVEL(mpidr, 2);
+ } else {
+ if (sccl_id)
+ *sccl_id = MPIDR_AFFINITY_LEVEL(mpidr, 2);
+ if (ccl_id)
+ *ccl_id = MPIDR_AFFINITY_LEVEL(mpidr, 1);
+ }
+}
+
+/*
+ * Check whether the CPU is associated with this uncore PMU
+ */
+static bool hisi_pmu_cpu_is_associated_pmu(struct hisi_pmu *hisi_pmu)
+{
+ int sccl_id, ccl_id;
+
+ if (hisi_pmu->ccl_id == -1) {
+ /* If CCL_ID is -1, the PMU only shares the same SCCL */
+ hisi_read_sccl_and_ccl_id(&sccl_id, NULL);
+
+ return sccl_id == hisi_pmu->sccl_id;
+ }
+
+ hisi_read_sccl_and_ccl_id(&sccl_id, &ccl_id);
+
+ return sccl_id == hisi_pmu->sccl_id && ccl_id == hisi_pmu->ccl_id;
+}
+
+int hisi_uncore_pmu_online_cpu(unsigned int cpu, struct hlist_node *node)
+{
+ struct hisi_pmu *hisi_pmu = hlist_entry_safe(node, struct hisi_pmu,
+ node);
+
+ if (!hisi_pmu_cpu_is_associated_pmu(hisi_pmu))
+ return 0;
+
+ cpumask_set_cpu(cpu, &hisi_pmu->associated_cpus);
+
+ /* If another CPU is already managing this PMU, simply return. */
+ if (hisi_pmu->on_cpu != -1)
+ return 0;
+
+ /* Use this CPU in cpumask for event counting */
+ hisi_pmu->on_cpu = cpu;
+
+ /* Overflow interrupt also should use the same CPU */
+ WARN_ON(irq_set_affinity(hisi_pmu->irq, cpumask_of(cpu)));
+
+ return 0;
+}
+
+int hisi_uncore_pmu_offline_cpu(unsigned int cpu, struct hlist_node *node)
+{
+ struct hisi_pmu *hisi_pmu = hlist_entry_safe(node, struct hisi_pmu,
+ node);
+ cpumask_t pmu_online_cpus;
+ unsigned int target;
+
+ if (!cpumask_test_and_clear_cpu(cpu, &hisi_pmu->associated_cpus))
+ return 0;
+
+ /* Nothing to do if this CPU doesn't own the PMU */
+ if (hisi_pmu->on_cpu != cpu)
+ return 0;
+
+ /* Give up ownership of the PMU */
+ hisi_pmu->on_cpu = -1;
+
+ /* Choose a new CPU to migrate ownership of the PMU to */
+ cpumask_and(&pmu_online_cpus, &hisi_pmu->associated_cpus,
+ cpu_online_mask);
+ target = cpumask_any_but(&pmu_online_cpus, cpu);
+ if (target >= nr_cpu_ids)
+ return 0;
+
+ perf_pmu_migrate_context(&hisi_pmu->pmu, cpu, target);
+ /* Use this CPU for event counting */
+ hisi_pmu->on_cpu = target;
+ WARN_ON(irq_set_affinity(hisi_pmu->irq, cpumask_of(target)));
+
+ return 0;
+}
diff --git a/drivers/perf/hisilicon/hisi_uncore_pmu.h b/drivers/perf/hisilicon/hisi_uncore_pmu.h
new file mode 100644
index 000000000000..f21226a0e9c6
--- /dev/null
+++ b/drivers/perf/hisilicon/hisi_uncore_pmu.h
@@ -0,0 +1,102 @@
+/*
+ * HiSilicon SoC Hardware event counters support
+ *
+ * Copyright (C) 2017 Hisilicon Limited
+ * Author: Anurup M <anurup.m@huawei.com>
+ * Shaokun Zhang <zhangshaokun@hisilicon.com>
+ *
+ * This code is based on the uncore PMUs like arm-cci and arm-ccn.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#ifndef __HISI_UNCORE_PMU_H__
+#define __HISI_UNCORE_PMU_H__
+
+#include <linux/cpumask.h>
+#include <linux/device.h>
+#include <linux/kernel.h>
+#include <linux/perf_event.h>
+#include <linux/types.h>
+
+#undef pr_fmt
+#define pr_fmt(fmt) "hisi_pmu: " fmt
+
+#define HISI_MAX_COUNTERS 0x10
+#define to_hisi_pmu(p) (container_of(p, struct hisi_pmu, pmu))
+
+#define HISI_PMU_ATTR(_name, _func, _config) \
+ (&((struct dev_ext_attribute[]) { \
+ { __ATTR(_name, 0444, _func, NULL), (void *)_config } \
+ })[0].attr.attr)
+
+#define HISI_PMU_FORMAT_ATTR(_name, _config) \
+ HISI_PMU_ATTR(_name, hisi_format_sysfs_show, (void *)_config)
+#define HISI_PMU_EVENT_ATTR(_name, _config) \
+ HISI_PMU_ATTR(_name, hisi_event_sysfs_show, (unsigned long)_config)
+
+struct hisi_pmu;
+
+struct hisi_uncore_ops {
+ void (*write_evtype)(struct hisi_pmu *, int, u32);
+ int (*get_event_idx)(struct perf_event *);
+ u64 (*read_counter)(struct hisi_pmu *, struct hw_perf_event *);
+ void (*write_counter)(struct hisi_pmu *, struct hw_perf_event *, u64);
+ void (*enable_counter)(struct hisi_pmu *, struct hw_perf_event *);
+ void (*disable_counter)(struct hisi_pmu *, struct hw_perf_event *);
+ void (*enable_counter_int)(struct hisi_pmu *, struct hw_perf_event *);
+ void (*disable_counter_int)(struct hisi_pmu *, struct hw_perf_event *);
+ void (*start_counters)(struct hisi_pmu *);
+ void (*stop_counters)(struct hisi_pmu *);
+};
+
+struct hisi_pmu_hwevents {
+ struct perf_event *hw_events[HISI_MAX_COUNTERS];
+ DECLARE_BITMAP(used_mask, HISI_MAX_COUNTERS);
+};
+
+/* Generic pmu struct for different pmu types */
+struct hisi_pmu {
+ struct pmu pmu;
+ const struct hisi_uncore_ops *ops;
+ struct hisi_pmu_hwevents pmu_events;
+ /* associated_cpus: All CPUs associated with the PMU */
+ cpumask_t associated_cpus;
+ /* CPU used for counting */
+ int on_cpu;
+ int irq;
+ struct device *dev;
+ struct hlist_node node;
+ int sccl_id;
+ int ccl_id;
+ void __iomem *base;
+ /* the ID of the PMU modules */
+ u32 index_id;
+ int num_counters;
+ int counter_bits;
+ /* check event code range */
+ int check_event;
+};
+
+int hisi_uncore_pmu_counter_valid(struct hisi_pmu *hisi_pmu, int idx);
+int hisi_uncore_pmu_get_event_idx(struct perf_event *event);
+void hisi_uncore_pmu_read(struct perf_event *event);
+int hisi_uncore_pmu_add(struct perf_event *event, int flags);
+void hisi_uncore_pmu_del(struct perf_event *event, int flags);
+void hisi_uncore_pmu_start(struct perf_event *event, int flags);
+void hisi_uncore_pmu_stop(struct perf_event *event, int flags);
+void hisi_uncore_pmu_set_event_period(struct perf_event *event);
+void hisi_uncore_pmu_event_update(struct perf_event *event);
+int hisi_uncore_pmu_event_init(struct perf_event *event);
+void hisi_uncore_pmu_enable(struct pmu *pmu);
+void hisi_uncore_pmu_disable(struct pmu *pmu);
+ssize_t hisi_event_sysfs_show(struct device *dev,
+ struct device_attribute *attr, char *buf);
+ssize_t hisi_format_sysfs_show(struct device *dev,
+ struct device_attribute *attr, char *buf);
+ssize_t hisi_cpumask_sysfs_show(struct device *dev,
+ struct device_attribute *attr, char *buf);
+int hisi_uncore_pmu_online_cpu(unsigned int cpu, struct hlist_node *node);
+int hisi_uncore_pmu_offline_cpu(unsigned int cpu, struct hlist_node *node);
+#endif /* __HISI_UNCORE_PMU_H__ */
diff --git a/drivers/perf/qcom_l2_pmu.c b/drivers/perf/qcom_l2_pmu.c
index b242cce10468..4fdc8486a8e4 100644
--- a/drivers/perf/qcom_l2_pmu.c
+++ b/drivers/perf/qcom_l2_pmu.c
@@ -92,6 +92,21 @@
#define reg_idx(reg, i) (((i) * IA_L2_REG_OFFSET) + reg##_BASE)
+/*
+ * Events
+ */
+#define L2_EVENT_CYCLES 0xfe
+#define L2_EVENT_DCACHE_OPS 0x400
+#define L2_EVENT_ICACHE_OPS 0x401
+#define L2_EVENT_TLBI 0x402
+#define L2_EVENT_BARRIERS 0x403
+#define L2_EVENT_TOTAL_READS 0x405
+#define L2_EVENT_TOTAL_WRITES 0x406
+#define L2_EVENT_TOTAL_REQUESTS 0x407
+#define L2_EVENT_LDREX 0x420
+#define L2_EVENT_STREX 0x421
+#define L2_EVENT_CLREX 0x422
+
static DEFINE_RAW_SPINLOCK(l2_access_lock);
/**
@@ -700,9 +715,12 @@ static struct attribute_group l2_cache_pmu_cpumask_group = {
/* CCG format for perf RAW codes. */
PMU_FORMAT_ATTR(l2_code, "config:4-11");
PMU_FORMAT_ATTR(l2_group, "config:0-3");
+PMU_FORMAT_ATTR(event, "config:0-11");
+
static struct attribute *l2_cache_pmu_formats[] = {
&format_attr_l2_code.attr,
&format_attr_l2_group.attr,
+ &format_attr_event.attr,
NULL,
};
@@ -711,9 +729,45 @@ static struct attribute_group l2_cache_pmu_format_group = {
.attrs = l2_cache_pmu_formats,
};
+static ssize_t l2cache_pmu_event_show(struct device *dev,
+ struct device_attribute *attr, char *page)
+{
+ struct perf_pmu_events_attr *pmu_attr;
+
+ pmu_attr = container_of(attr, struct perf_pmu_events_attr, attr);
+ return sprintf(page, "event=0x%02llx\n", pmu_attr->id);
+}
+
+#define L2CACHE_EVENT_ATTR(_name, _id) \
+ (&((struct perf_pmu_events_attr[]) { \
+ { .attr = __ATTR(_name, 0444, l2cache_pmu_event_show, NULL), \
+ .id = _id, } \
+ })[0].attr.attr)
+
+static struct attribute *l2_cache_pmu_events[] = {
+ L2CACHE_EVENT_ATTR(cycles, L2_EVENT_CYCLES),
+ L2CACHE_EVENT_ATTR(dcache-ops, L2_EVENT_DCACHE_OPS),
+ L2CACHE_EVENT_ATTR(icache-ops, L2_EVENT_ICACHE_OPS),
+ L2CACHE_EVENT_ATTR(tlbi, L2_EVENT_TLBI),
+ L2CACHE_EVENT_ATTR(barriers, L2_EVENT_BARRIERS),
+ L2CACHE_EVENT_ATTR(total-reads, L2_EVENT_TOTAL_READS),
+ L2CACHE_EVENT_ATTR(total-writes, L2_EVENT_TOTAL_WRITES),
+ L2CACHE_EVENT_ATTR(total-requests, L2_EVENT_TOTAL_REQUESTS),
+ L2CACHE_EVENT_ATTR(ldrex, L2_EVENT_LDREX),
+ L2CACHE_EVENT_ATTR(strex, L2_EVENT_STREX),
+ L2CACHE_EVENT_ATTR(clrex, L2_EVENT_CLREX),
+ NULL
+};
+
+static struct attribute_group l2_cache_pmu_events_group = {
+ .name = "events",
+ .attrs = l2_cache_pmu_events,
+};
+
static const struct attribute_group *l2_cache_pmu_attr_grps[] = {
&l2_cache_pmu_format_group,
&l2_cache_pmu_cpumask_group,
+ &l2_cache_pmu_events_group,
NULL,
};