summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorThomas Richter <tmricht@linux.ibm.com>2025-11-05 15:38:59 +0100
committerHeiko Carstens <hca@linux.ibm.com>2025-11-14 11:30:07 +0100
commit35a27bad075d594ebaabc62f94b665fa2a135433 (patch)
treec1cd03f6a4db52a75e5146b7601e7ab8a04e9b2a
parentf12473541356f5765e4130a7b5af77862a872a51 (diff)
s390/pai_crypto: Make pai_root per-PMU and unify naming
Prepare the common PAI PMU driver to handle multiple PMUs. Convert pai_root into an array indexed by PAI_PMU_IDX(event) so that per-CPU state becomes per-PMU. Adjust all call sites accordingly. Rename KMSG_COMPONENT and the s390dbf buffer from "pai_crypto" to "pai" for consistent naming. No functional change intended beyond log identifiers. Signed-off-by: Thomas Richter <tmricht@linux.ibm.com> Reviewed-by: Jan Polensky <japo@linux.ibm.com> Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
-rw-r--r--arch/s390/kernel/perf_pai_crypto.c71
1 files changed, 36 insertions, 35 deletions
diff --git a/arch/s390/kernel/perf_pai_crypto.c b/arch/s390/kernel/perf_pai_crypto.c
index 870b84b0ce5f..94551b3d09f3 100644
--- a/arch/s390/kernel/perf_pai_crypto.c
+++ b/arch/s390/kernel/perf_pai_crypto.c
@@ -2,10 +2,10 @@
/*
* Performance event support - Processor Activity Instrumentation Facility
*
- * Copyright IBM Corp. 2022
+ * Copyright IBM Corp. 2026
* Author(s): Thomas Richter <tmricht@linux.ibm.com>
*/
-#define KMSG_COMPONENT "pai_crypto"
+#define KMSG_COMPONENT "pai"
#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt
#include <linux/kernel.h>
@@ -49,7 +49,7 @@ struct pai_mapptr {
static struct pai_root { /* Anchor to per CPU data */
refcount_t refcnt; /* Overall active events */
struct pai_mapptr __percpu *mapptr;
-} pai_root;
+} pai_root[PAI_PMU_MAX];
/* This table defines the different parameters of the PAI PMUs. During
* initialization the machine dependent values are extracted and saved.
@@ -74,14 +74,14 @@ struct pai_pmu { /* Define PAI PMU characteristics */
static struct pai_pmu pai_pmu[]; /* Forward declaration */
/* Free per CPU data when the last event is removed. */
-static void pai_root_free(void)
+static void pai_root_free(int idx)
{
- if (refcount_dec_and_test(&pai_root.refcnt)) {
- free_percpu(pai_root.mapptr);
- pai_root.mapptr = NULL;
+ if (refcount_dec_and_test(&pai_root[idx].refcnt)) {
+ free_percpu(pai_root[idx].mapptr);
+ pai_root[idx].mapptr = NULL;
}
- debug_sprintf_event(paidbg, 5, "%s root.refcount %d\n", __func__,
- refcount_read(&pai_root.refcnt));
+ debug_sprintf_event(paidbg, 5, "%s root[%d].refcount %d\n", __func__,
+ idx, refcount_read(&pai_root[idx].refcnt));
}
/*
@@ -90,14 +90,14 @@ static void pai_root_free(void)
* CPUs possible, which might be larger than the number of CPUs currently
* online.
*/
-static int pai_root_alloc(void)
+static int pai_root_alloc(int idx)
{
- if (!refcount_inc_not_zero(&pai_root.refcnt)) {
+ if (!refcount_inc_not_zero(&pai_root[idx].refcnt)) {
/* The memory is already zeroed. */
- pai_root.mapptr = alloc_percpu(struct pai_mapptr);
- if (!pai_root.mapptr)
+ pai_root[idx].mapptr = alloc_percpu(struct pai_mapptr);
+ if (!pai_root[idx].mapptr)
return -ENOMEM;
- refcount_set(&pai_root.refcnt, 1);
+ refcount_set(&pai_root[idx].refcnt, 1);
}
return 0;
}
@@ -119,17 +119,18 @@ static void pai_free(struct pai_mapptr *mp)
*/
static void pai_event_destroy_cpu(struct perf_event *event, int cpu)
{
- struct pai_mapptr *mp = per_cpu_ptr(pai_root.mapptr, cpu);
+ int idx = PAI_PMU_IDX(event);
+ struct pai_mapptr *mp = per_cpu_ptr(pai_root[idx].mapptr, cpu);
struct pai_map *cpump = mp->mapptr;
mutex_lock(&pai_reserve_mutex);
- debug_sprintf_event(paidbg, 5, "%s event %#llx cpu %d users %d "
- "refcnt %u\n", __func__, event->attr.config,
+ debug_sprintf_event(paidbg, 5, "%s event %#llx idx %d cpu %d users %d "
+ "refcnt %u\n", __func__, event->attr.config, idx,
event->cpu, cpump->active_events,
refcount_read(&cpump->refcnt));
if (refcount_dec_and_test(&cpump->refcnt))
pai_free(mp);
- pai_root_free();
+ pai_root_free(idx);
mutex_unlock(&pai_reserve_mutex);
}
@@ -162,10 +163,10 @@ static u64 pai_getctr(unsigned long *page, int nr, unsigned long offset)
*/
static u64 pai_getdata(struct perf_event *event, bool kernel)
{
- struct pai_mapptr *mp = this_cpu_ptr(pai_root.mapptr);
- struct pai_map *cpump = mp->mapptr;
int idx = PAI_PMU_IDX(event);
+ struct pai_mapptr *mp = this_cpu_ptr(pai_root[idx].mapptr);
struct pai_pmu *pp = &pai_pmu[idx];
+ struct pai_map *cpump = mp->mapptr;
unsigned int i;
u64 sum = 0;
@@ -213,12 +214,12 @@ static int pai_alloc_cpu(struct perf_event *event, int cpu)
mutex_lock(&pai_reserve_mutex);
/* Allocate root node */
- rc = pai_root_alloc();
+ rc = pai_root_alloc(idx);
if (rc)
goto unlock;
/* Allocate node for this event */
- mp = per_cpu_ptr(pai_root.mapptr, cpu);
+ mp = per_cpu_ptr(pai_root[idx].mapptr, cpu);
cpump = mp->mapptr;
if (!cpump) { /* Paicrypt_map allocated? */
rc = -ENOMEM;
@@ -250,7 +251,7 @@ undo:
* the event in not created, its destroy() function is never
* invoked. Adjust the reference counter for the anchor.
*/
- pai_root_free();
+ pai_root_free(idx);
}
unlock:
mutex_unlock(&pai_reserve_mutex);
@@ -387,7 +388,7 @@ static void pai_start(struct perf_event *event, int flags,
{
int idx = PAI_PMU_IDX(event);
struct pai_pmu *pp = &pai_pmu[idx];
- struct pai_mapptr *mp = this_cpu_ptr(pai_root.mapptr);
+ struct pai_mapptr *mp = this_cpu_ptr(pai_root[idx].mapptr);
struct pai_map *cpump = mp->mapptr;
u64 sum;
@@ -413,9 +414,9 @@ static void paicrypt_start(struct perf_event *event, int flags)
static int pai_add(struct perf_event *event, int flags)
{
- struct pai_mapptr *mp = this_cpu_ptr(pai_root.mapptr);
- struct pai_map *cpump = mp->mapptr;
int idx = PAI_PMU_IDX(event);
+ struct pai_mapptr *mp = this_cpu_ptr(pai_root[idx].mapptr);
+ struct pai_map *cpump = mp->mapptr;
unsigned long ccd;
if (++cpump->active_events == 1) {
@@ -437,9 +438,9 @@ static int paicrypt_add(struct perf_event *event, int flags)
static void pai_have_sample(struct perf_event *, struct pai_map *);
static void pai_stop(struct perf_event *event, int flags)
{
- struct pai_mapptr *mp = this_cpu_ptr(pai_root.mapptr);
- struct pai_map *cpump = mp->mapptr;
int idx = PAI_PMU_IDX(event);
+ struct pai_mapptr *mp = this_cpu_ptr(pai_root[idx].mapptr);
+ struct pai_map *cpump = mp->mapptr;
if (!event->attr.sample_period) { /* Counting */
pai_pmu[idx].pmu->read(event);
@@ -462,9 +463,9 @@ static void paicrypt_stop(struct perf_event *event, int flags)
static void pai_del(struct perf_event *event, int flags)
{
- struct pai_mapptr *mp = this_cpu_ptr(pai_root.mapptr);
- struct pai_map *cpump = mp->mapptr;
int idx = PAI_PMU_IDX(event);
+ struct pai_mapptr *mp = this_cpu_ptr(pai_root[idx].mapptr);
+ struct pai_map *cpump = mp->mapptr;
pai_pmu[idx].pmu->stop(event, PERF_EF_UPDATE);
if (--cpump->active_events == 0) {
@@ -587,9 +588,9 @@ static void pai_have_sample(struct perf_event *event, struct pai_map *cpump)
}
/* Check if there is data to be saved on schedule out of a task. */
-static void pai_have_samples(void)
+static void pai_have_samples(int idx)
{
- struct pai_mapptr *mp = this_cpu_ptr(pai_root.mapptr);
+ struct pai_mapptr *mp = this_cpu_ptr(pai_root[idx].mapptr);
struct pai_map *cpump = mp->mapptr;
struct perf_event *event;
@@ -607,7 +608,7 @@ static void paicrypt_sched_task(struct perf_event_pmu_context *pmu_ctx,
* results on schedule_out and if page was dirty, save old values.
*/
if (!sched_in)
- pai_have_samples();
+ pai_have_samples(PAI_PMU_CRYPTO);
}
/* Attribute definitions for paicrypt interface. As with other CPU
@@ -982,7 +983,7 @@ static int __init paipmu_setup(void)
return install_ok;
}
-static int __init paicrypt_init(void)
+static int __init pai_init(void)
{
/* Setup s390dbf facility */
paidbg = debug_register(KMSG_COMPONENT, 32, 256, 128);
@@ -1001,4 +1002,4 @@ static int __init paicrypt_init(void)
return 0;
}
-device_initcall(paicrypt_init);
+device_initcall(pai_init);