perf: Rework the PMU methods

Replace pmu::{enable,disable,start,stop,unthrottle} with
pmu::{add,del,start,stop}, all of which take a flags argument.

The new interface extends the capability to stop a counter while
keeping it scheduled on the PMU. We replace the throttled state with
the generic stopped state.

This also allows us to efficiently stop/start counters over certain
code paths (like IRQ handlers).

It also allows scheduling a counter without it starting, allowing for
a generic frozen state (useful for rotating stopped counters).

The stopped state is implemented in two different ways, depending on
how the architecture implemented the throttled state:

 1) We disable the counter:
    a) the pmu has per-counter enable bits, we flip that
    b) we program a NOP event, preserving the counter state

 2) We store the counter state and ignore all read/overflow events

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: paulus <paulus@samba.org>
Cc: stephane eranian <eranian@googlemail.com>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Lin Ming <ming.m.lin@intel.com>
Cc: Yanmin <yanmin_zhang@linux.intel.com>
Cc: Deng-Cheng Zhu <dengcheng.zhu@gmail.com>
Cc: David Miller <davem@davemloft.net>
Cc: Michael Cree <mcree@orcon.net.nz>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

3 files changed, 63 insertions, 47 deletions

diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index 79705ac45019..dd6fec710677 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -583,7 +583,7 @@ static void x86_pmu_disable_all(void)
         }
 }
 
-static void x86_pmu_pmu_disable(struct pmu *pmu)
+static void x86_pmu_disable(struct pmu *pmu)
 {
         struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
 
@@ -800,10 +800,10 @@ static inline int match_prev_assignment(struct hw_perf_event *hwc,
                 hwc->last_tag == cpuc->tags[i];
 }
 
-static int x86_pmu_start(struct perf_event *event);
-static void x86_pmu_stop(struct perf_event *event);
+static void x86_pmu_start(struct perf_event *event, int flags);
+static void x86_pmu_stop(struct perf_event *event, int flags);
 
-static void x86_pmu_pmu_enable(struct pmu *pmu)
+static void x86_pmu_enable(struct pmu *pmu)
 {
         struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
         struct perf_event *event;
@@ -839,7 +839,14 @@ static void x86_pmu_pmu_enable(struct pmu *pmu)
                             match_prev_assignment(hwc, cpuc, i))
                                 continue;
 
-                        x86_pmu_stop(event);
+                        /*
+                         * Ensure we don't accidentally enable a stopped
+                         * counter simply because we rescheduled.
+                         */
+                        if (hwc->state & PERF_HES_STOPPED)
+                                hwc->state |= PERF_HES_ARCH;
+
+                        x86_pmu_stop(event, PERF_EF_UPDATE);
                 }
 
                 for (i = 0; i < cpuc->n_events; i++) {
@@ -851,7 +858,10 @@ static void x86_pmu_pmu_enable(struct pmu *pmu)
                         else if (i < n_running)
                                 continue;
 
-                        x86_pmu_start(event);
+                        if (hwc->state & PERF_HES_ARCH)
+                                continue;
+
+                        x86_pmu_start(event, PERF_EF_RELOAD);
                 }
                 cpuc->n_added = 0;
                 perf_events_lapic_init();
@@ -952,15 +962,12 @@ static void x86_pmu_enable_event(struct perf_event *event)
 }
 
 /*
- * activate a single event
+ * Add a single event to the PMU.
  *
  * The event is added to the group of enabled events
  * but only if it can be scehduled with existing events.
- *
- * Called with PMU disabled. If successful and return value 1,
- * then guaranteed to call perf_enable() and hw_perf_enable()
  */
-static int x86_pmu_enable(struct perf_event *event)
+static int x86_pmu_add(struct perf_event *event, int flags)
 {
         struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
         struct hw_perf_event *hwc;
@@ -975,10 +982,14 @@ static int x86_pmu_enable(struct perf_event *event)
         if (ret < 0)
                 goto out;
 
+        hwc->state = PERF_HES_UPTODATE | PERF_HES_STOPPED;
+        if (!(flags & PERF_EF_START))
+                hwc->state |= PERF_HES_ARCH;
+
         /*
          * If group events scheduling transaction was started,
          * skip the schedulability test here, it will be peformed
-         * at commit time(->commit_txn) as a whole
+         * at commit time (->commit_txn) as a whole
          */
         if (cpuc->group_flag & PERF_EVENT_TXN)
                 goto done_collect;
@@ -1003,27 +1014,28 @@ out:
         return ret;
 }
 
-static int x86_pmu_start(struct perf_event *event)
+static void x86_pmu_start(struct perf_event *event, int flags)
 {
         struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
         int idx = event->hw.idx;
 
-        if (idx == -1)
-                return -EAGAIN;
+        if (WARN_ON_ONCE(!(event->hw.state & PERF_HES_STOPPED)))
+                return;
+
+        if (WARN_ON_ONCE(idx == -1))
+                return;
+
+        if (flags & PERF_EF_RELOAD) {
+                WARN_ON_ONCE(!(event->hw.state & PERF_HES_UPTODATE));
+                x86_perf_event_set_period(event);
+        }
+
+        event->hw.state = 0;
 
-        x86_perf_event_set_period(event);
         cpuc->events[idx] = event;
         __set_bit(idx, cpuc->active_mask);
         x86_pmu.enable(event);
         perf_event_update_userpage(event);
-
-        return 0;
-}
-
-static void x86_pmu_unthrottle(struct perf_event *event)
-{
-        int ret = x86_pmu_start(event);
-        WARN_ON_ONCE(ret);
 }
 
 void perf_event_print_debug(void)
@@ -1080,27 +1092,29 @@ void perf_event_print_debug(void)
         local_irq_restore(flags);
 }
 
-static void x86_pmu_stop(struct perf_event *event)
+static void x86_pmu_stop(struct perf_event *event, int flags)
 {
         struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
         struct hw_perf_event *hwc = &event->hw;
-        int idx = hwc->idx;
-
-        if (!__test_and_clear_bit(idx, cpuc->active_mask))
-                return;
-
-        x86_pmu.disable(event);
 
-        /*
-         * Drain the remaining delta count out of a event
-         * that we are disabling:
-         */
-        x86_perf_event_update(event);
+        if (__test_and_clear_bit(hwc->idx, cpuc->active_mask)) {
+                x86_pmu.disable(event);
+                cpuc->events[hwc->idx] = NULL;
+                WARN_ON_ONCE(hwc->state & PERF_HES_STOPPED);
+                hwc->state |= PERF_HES_STOPPED;
+        }
 
-        cpuc->events[idx] = NULL;
+        if ((flags & PERF_EF_UPDATE) && !(hwc->state & PERF_HES_UPTODATE)) {
+                /*
+                 * Drain the remaining delta count out of a event
+                 * that we are disabling:
+                 */
+                x86_perf_event_update(event);
+                hwc->state |= PERF_HES_UPTODATE;
+        }
 }
 
-static void x86_pmu_disable(struct perf_event *event)
+static void x86_pmu_del(struct perf_event *event, int flags)
 {
         struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
         int i;
@@ -1113,7 +1127,7 @@ static void x86_pmu_disable(struct perf_event *event)
         if (cpuc->group_flag & PERF_EVENT_TXN)
                 return;
 
-        x86_pmu_stop(event);
+        x86_pmu_stop(event, PERF_EF_UPDATE);
 
         for (i = 0; i < cpuc->n_events; i++) {
                 if (event == cpuc->event_list[i]) {
@@ -1165,7 +1179,7 @@ static int x86_pmu_handle_irq(struct pt_regs *regs)
                         continue;
 
                 if (perf_event_overflow(event, 1, &data, regs))
-                        x86_pmu_stop(event);
+                        x86_pmu_stop(event, 0);
         }
 
         if (handled)
@@ -1605,15 +1619,17 @@ int x86_pmu_event_init(struct perf_event *event)
 }
 
 static struct pmu pmu = {
-        .pmu_enable     = x86_pmu_pmu_enable,
-        .pmu_disable    = x86_pmu_pmu_disable,
+        .pmu_enable     = x86_pmu_enable,
+        .pmu_disable    = x86_pmu_disable,
+
         .event_init     = x86_pmu_event_init,
-        .enable         = x86_pmu_enable,
-        .disable        = x86_pmu_disable,
+
+        .add            = x86_pmu_add,
+        .del            = x86_pmu_del,
         .start          = x86_pmu_start,
         .stop           = x86_pmu_stop,
         .read           = x86_pmu_read,
-        .unthrottle     = x86_pmu_unthrottle,
+
         .start_txn      = x86_pmu_start_txn,
         .cancel_txn     = x86_pmu_cancel_txn,
         .commit_txn     = x86_pmu_commit_txn,
diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c
index ee05c90012d2..82395f2378ec 100644
--- a/arch/x86/kernel/cpu/perf_event_intel.c
+++ b/arch/x86/kernel/cpu/perf_event_intel.c
@@ -763,7 +763,7 @@ again:
                 data.period = event->hw.last_period;
 
                 if (perf_event_overflow(event, 1, &data, regs))
-                        x86_pmu_stop(event);
+                        x86_pmu_stop(event, 0);
         }
 
         /*
diff --git a/arch/x86/kernel/cpu/perf_event_intel_ds.c b/arch/x86/kernel/cpu/perf_event_intel_ds.c
index 18018d1311cd..9893a2f77b7a 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_ds.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_ds.c
@@ -491,7 +491,7 @@ static void __intel_pmu_pebs_event(struct perf_event *event,
                 regs.flags &= ~PERF_EFLAGS_EXACT;
 
         if (perf_event_overflow(event, 1, &data, &regs))
-                x86_pmu_stop(event);
+                x86_pmu_stop(event, 0);
 }
 
 static void intel_pmu_drain_pebs_core(struct pt_regs *iregs)