Re: [PATCH 4/4] perf, amd: Enable northbridge performance counterson AMD family 15h
From: Jacob Shin
Date: Fri Nov 16 2012 - 17:12:47 EST
On Fri, Nov 16, 2012 at 08:32:24PM +0100, Robert Richter wrote:
> On 16.11.12 13:00:30, Jacob Shin wrote:
> > On Fri, Nov 16, 2012 at 07:43:44PM +0100, Robert Richter wrote:
> > > On 15.11.12 15:31:53, Jacob Shin wrote:
> > > > @@ -156,31 +161,28 @@ static inline int amd_pmu_addr_offset(int index)
> > > > if (offset)
> > > > return offset;
> > > >
> > > > - if (!cpu_has_perfctr_core)
> > > > + if (!cpu_has_perfctr_core) {
> > > > offset = index;
> > > > - else
> > > > + ncore = AMD64_NUM_COUNTERS;
> > > > + } else {
>
> First calculation:
>
> > > > offset = index << 1;
> > > > + ncore = AMD64_NUM_COUNTERS_CORE;
> > > > + }
> > > > +
> > > > + /* find offset of NB counters with respect to x86_pmu.eventsel */
> > > > + if (cpu_has_perfctr_nb) {
> > > > + if (index >= ncore && index < (ncore + AMD64_NUM_COUNTERS_NB))
>
> Second calculation:
>
> > > > + offset = (MSR_F15H_NB_PERF_CTL - x86_pmu.eventsel) +
> > > > + ((index - ncore) << 1);
> > > > + }
> > >
> > > There is duplicate calculatoin of offset in some cases. Better we
> > > avoid this.
> >
> > Which cases? The code calculates the offset for a given index the very
> > first time it is called, stores it, and uses that stored offset from
> > then on. My [PATCH 3/4] sets that up.
>
> One case above.
>
> It looks like the paths should be defined more clearly.
Per comments above the function, I was logically going down the cases,
1. is this a legacy counter?
2. is this a perfctr_core counter?
3. is this a perfctr_nb counter?
To me it seems clear ..
>
> > > > @@ -323,6 +368,16 @@ __amd_get_nb_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *ev
> > > > if (new == -1)
> > > > return &emptyconstraint;
> > > >
> > > > + /* set up interrupts to be delivered only to this core */
> > > > + if (cpu_has_perfctr_nb) {
> > > > + struct cpuinfo_x86 *c = &cpu_data(smp_processor_id());
> > > > +
> > > > + hwc->config |= AMD_PERFMON_EVENTSEL_INT_CORE_ENABLE;
> > > > + hwc->config &= ~AMD_PERFMON_EVENTSEL_INT_CORE_SEL_MASK;
> > > > + hwc->config |= (0ULL | (c->cpu_core_id)) <<
> > > > + AMD_PERFMON_EVENTSEL_INT_CORE_SEL_SHIFT;
> > > > + }
> > >
> > > Looks like a hack to me. The constaints handler is only supposed to
> > > determine constraints and not to touch anything in the event's
> > > structure. This should be done later when setting up hwc->config in
> > > amd_nb_event_config() or so.
> >
> > Hm.. is the hwc->config called after constraints have been set up
> > already? If so, I'll change it ..
>
> Should be, since the hw register can be setup only after the counter
> is selected.
>
> >
> > >
> > > I also do not think that smp_processor_id() is the right thing to do
> > > here. Since cpu_hw_events is per-cpu the cpu is already selected.
> >
> > Yeah, I could not figure out how to get the cpu number from cpuc. Is
> > there a container_of kind of thing that I can do to get the cpu number
> > ?
>
> At some point event->cpu is assigned, I think.
Great, thanks for this hint!
So here is v3, how does this look? If okay, could you add Reviewed-by or
Acked-by ? After that, I'll send out the patchbomb again with review/ack
on patch [3/4] and [4/4]
diff --git a/arch/x86/include/asm/cpufeature.h b/arch/x86/include/asm/cpufeature.h
index 8c297aa..b05c722 100644
--- a/arch/x86/include/asm/cpufeature.h
+++ b/arch/x86/include/asm/cpufeature.h
@@ -167,6 +167,7 @@
#define X86_FEATURE_TBM (6*32+21) /* trailing bit manipulations */
#define X86_FEATURE_TOPOEXT (6*32+22) /* topology extensions CPUID leafs */
#define X86_FEATURE_PERFCTR_CORE (6*32+23) /* core performance counter extensions */
+#define X86_FEATURE_PERFCTR_NB (6*32+24) /* core performance counter extensions */
/*
* Auxiliary flags: Linux defined - For features scattered in various
@@ -308,6 +309,7 @@ extern const char * const x86_power_flags[32];
#define cpu_has_hypervisor boot_cpu_has(X86_FEATURE_HYPERVISOR)
#define cpu_has_pclmulqdq boot_cpu_has(X86_FEATURE_PCLMULQDQ)
#define cpu_has_perfctr_core boot_cpu_has(X86_FEATURE_PERFCTR_CORE)
+#define cpu_has_perfctr_nb boot_cpu_has(X86_FEATURE_PERFCTR_NB)
#define cpu_has_cx8 boot_cpu_has(X86_FEATURE_CX8)
#define cpu_has_cx16 boot_cpu_has(X86_FEATURE_CX16)
#define cpu_has_eager_fpu boot_cpu_has(X86_FEATURE_EAGER_FPU)
diff --git a/arch/x86/include/asm/msr-index.h b/arch/x86/include/asm/msr-index.h
index 7f0edce..e67ff1e 100644
--- a/arch/x86/include/asm/msr-index.h
+++ b/arch/x86/include/asm/msr-index.h
@@ -157,6 +157,8 @@
/* Fam 15h MSRs */
#define MSR_F15H_PERF_CTL 0xc0010200
#define MSR_F15H_PERF_CTR 0xc0010201
+#define MSR_F15H_NB_PERF_CTL 0xc0010240
+#define MSR_F15H_NB_PERF_CTR 0xc0010241
/* Fam 10h MSRs */
#define MSR_FAM10H_MMIO_CONF_BASE 0xc0010058
diff --git a/arch/x86/include/asm/perf_event.h b/arch/x86/include/asm/perf_event.h
index 4fabcdf..df97186 100644
--- a/arch/x86/include/asm/perf_event.h
+++ b/arch/x86/include/asm/perf_event.h
@@ -29,9 +29,14 @@
#define ARCH_PERFMON_EVENTSEL_INV (1ULL << 23)
#define ARCH_PERFMON_EVENTSEL_CMASK 0xFF000000ULL
+#define AMD_PERFMON_EVENTSEL_INT_CORE_ENABLE (1ULL << 36)
#define AMD_PERFMON_EVENTSEL_GUESTONLY (1ULL << 40)
#define AMD_PERFMON_EVENTSEL_HOSTONLY (1ULL << 41)
+#define AMD_PERFMON_EVENTSEL_INT_CORE_SEL_SHIFT 37
+#define AMD_PERFMON_EVENTSEL_INT_CORE_SEL_MASK \
+ (0xFULL << AMD_PERFMON_EVENTSEL_INT_CORE_SEL_SHIFT)
+
#define AMD64_EVENTSEL_EVENT \
(ARCH_PERFMON_EVENTSEL_EVENT | (0x0FULL << 32))
#define INTEL_ARCH_EVENT_MASK \
@@ -46,8 +51,12 @@
#define AMD64_RAW_EVENT_MASK \
(X86_RAW_EVENT_MASK | \
AMD64_EVENTSEL_EVENT)
+#define AMD64_NB_EVENT_MASK \
+ (AMD64_EVENTSEL_EVENT | \
+ ARCH_PERFMON_EVENTSEL_UMASK)
#define AMD64_NUM_COUNTERS 4
#define AMD64_NUM_COUNTERS_CORE 6
+#define AMD64_NUM_COUNTERS_NB 4
#define ARCH_PERFMON_UNHALTED_CORE_CYCLES_SEL 0x3c
#define ARCH_PERFMON_UNHALTED_CORE_CYCLES_UMASK (0x00 << 8)
diff --git a/arch/x86/kernel/cpu/perf_event_amd.c b/arch/x86/kernel/cpu/perf_event_amd.c
index d6e3337..80ad803 100644
--- a/arch/x86/kernel/cpu/perf_event_amd.c
+++ b/arch/x86/kernel/cpu/perf_event_amd.c
@@ -132,6 +132,8 @@ static u64 amd_pmu_event_map(int hw_event)
return amd_perfmon_event_map[hw_event];
}
+static struct event_constraint *amd_nb_event_constraint;
+
/*
* Previously calculated offsets
*/
@@ -143,10 +145,15 @@ static unsigned int addr_offsets[X86_PMC_IDX_MAX] __read_mostly;
*
* CPUs with core performance counter extensions:
* 6 counters starting at 0xc0010200 each offset by 2
+ *
+ * CPUs with north bridge performance counter extensions:
+ * 4 additional counters starting at 0xc0010240 each offset by 2
+ * (indexed right above either one of the above core counters)
*/
static inline int amd_pmu_addr_offset(int index)
{
int offset;
+ int ncore;
if (!index)
return index;
@@ -156,31 +163,27 @@ static inline int amd_pmu_addr_offset(int index)
if (offset)
return offset;
- if (!cpu_has_perfctr_core)
+ if (!cpu_has_perfctr_core) {
offset = index;
- else
+ ncore = AMD64_NUM_COUNTERS;
+ } else {
offset = index << 1;
+ ncore = AMD64_NUM_COUNTERS_CORE;
+ }
+
+ /* find offset of NB counters with respect to x86_pmu.eventsel */
+ if (amd_nb_event_constraint &&
+ test_bit(index, amd_nb_event_constraint->idxmsk))
+ offset = (MSR_F15H_NB_PERF_CTL - x86_pmu.eventsel) +
+ ((index - ncore) << 1);
addr_offsets[index] = offset;
return offset;
}
-static int amd_pmu_hw_config(struct perf_event *event)
+static int __amd_core_hw_config(struct perf_event *event)
{
- int ret;
-
- /* pass precise event sampling to ibs: */
- if (event->attr.precise_ip && get_ibs_caps())
- return -ENOENT;
-
- ret = x86_pmu_hw_config(event);
- if (ret)
- return ret;
-
- if (has_branch_stack(event))
- return -EOPNOTSUPP;
-
if (event->attr.exclude_host && event->attr.exclude_guest)
/*
* When HO == GO == 1 the hardware treats that as GO == HO == 0
@@ -194,10 +197,41 @@ static int amd_pmu_hw_config(struct perf_event *event)
else if (event->attr.exclude_guest)
event->hw.config |= AMD_PERFMON_EVENTSEL_HOSTONLY;
- if (event->attr.type != PERF_TYPE_RAW)
- return 0;
+ return 0;
+}
+
+/*
+ * NB counters do not support the following event select bits:
+ * Host/Guest only
+ * Counter mask
+ * Invert counter mask
+ * Edge detect
+ * OS/User mode
+ */
+static int __amd_nb_hw_config(struct perf_event *event)
+{
+ if (event->attr.exclude_user || event->attr.exclude_kernel ||
+ event->attr.exclude_host || event->attr.exclude_guest)
+ return -EINVAL;
+
+ /* set up interrupts to be delivered only to this core */
+ if (cpu_has_perfctr_nb) {
+ struct cpuinfo_x86 *c = &cpu_data(event->cpu);
+
+ event->hw.config |= AMD_PERFMON_EVENTSEL_INT_CORE_ENABLE;
+ event->hw.config &= ~AMD_PERFMON_EVENTSEL_INT_CORE_SEL_MASK;
+ event->hw.config |= (0ULL | (c->cpu_core_id)) <<
+ AMD_PERFMON_EVENTSEL_INT_CORE_SEL_SHIFT;
+ }
+
+ event->hw.config &= ~(ARCH_PERFMON_EVENTSEL_USR |
+ ARCH_PERFMON_EVENTSEL_OS);
- event->hw.config |= event->attr.config & AMD64_RAW_EVENT_MASK;
+ if (event->hw.config & ~(AMD64_NB_EVENT_MASK |
+ ARCH_PERFMON_EVENTSEL_INT |
+ AMD_PERFMON_EVENTSEL_INT_CORE_ENABLE |
+ AMD_PERFMON_EVENTSEL_INT_CORE_SEL_MASK))
+ return -EINVAL;
return 0;
}
@@ -215,6 +249,11 @@ static inline int amd_is_nb_event(struct hw_perf_event *hwc)
return (hwc->config & 0xe0) == 0xe0;
}
+static inline int amd_is_perfctr_nb_event(struct hw_perf_event *hwc)
+{
+ return amd_nb_event_constraint && amd_is_nb_event(hwc);
+}
+
static inline int amd_has_nb(struct cpu_hw_events *cpuc)
{
struct amd_nb *nb = cpuc->amd_nb;
@@ -222,6 +261,30 @@ static inline int amd_has_nb(struct cpu_hw_events *cpuc)
return nb && nb->nb_id != -1;
}
+static int amd_pmu_hw_config(struct perf_event *event)
+{
+ int ret;
+
+ /* pass precise event sampling to ibs: */
+ if (event->attr.precise_ip && get_ibs_caps())
+ return -ENOENT;
+
+ if (has_branch_stack(event))
+ return -EOPNOTSUPP;
+
+ ret = x86_pmu_hw_config(event);
+ if (ret)
+ return ret;
+
+ if (event->attr.type == PERF_TYPE_RAW)
+ event->hw.config |= event->attr.config & AMD64_RAW_EVENT_MASK;
+
+ if (amd_is_perfctr_nb_event(&event->hw))
+ return __amd_nb_hw_config(event);
+
+ return __amd_core_hw_config(event);
+}
+
static void __amd_put_nb_event_constraints(struct cpu_hw_events *cpuc,
struct perf_event *event)
{
@@ -422,7 +485,10 @@ amd_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event)
if (!(amd_has_nb(cpuc) && amd_is_nb_event(&event->hw)))
return &unconstrained;
- return __amd_get_nb_event_constraints(cpuc, event, &unconstrained);
+ return __amd_get_nb_event_constraints(cpuc, event,
+ amd_nb_event_constraint ?
+ amd_nb_event_constraint :
+ &unconstrained);
}
static void amd_put_event_constraints(struct cpu_hw_events *cpuc,
@@ -521,6 +587,9 @@ static struct event_constraint amd_f15_PMC30 = EVENT_CONSTRAINT_OVERLAP(0, 0x09,
static struct event_constraint amd_f15_PMC50 = EVENT_CONSTRAINT(0, 0x3F, 0);
static struct event_constraint amd_f15_PMC53 = EVENT_CONSTRAINT(0, 0x38, 0);
+static struct event_constraint amd_NBPMC96 = EVENT_CONSTRAINT(0, 0x3C0, 0);
+static struct event_constraint amd_NBPMC74 = EVENT_CONSTRAINT(0, 0xF0, 0);
+
static struct event_constraint *
amd_get_event_constraints_f15h(struct cpu_hw_events *cpuc, struct perf_event *event)
{
@@ -586,8 +655,11 @@ amd_get_event_constraints_f15h(struct cpu_hw_events *cpuc, struct perf_event *ev
return &amd_f15_PMC20;
}
case AMD_EVENT_NB:
- /* not yet implemented */
- return &emptyconstraint;
+ if (cpuc->is_fake)
+ return amd_nb_event_constraint;
+
+ return __amd_get_nb_event_constraints(cpuc, event,
+ amd_nb_event_constraint);
default:
return &emptyconstraint;
}
@@ -625,7 +697,7 @@ static __initconst const struct x86_pmu amd_pmu = {
static int setup_event_constraints(void)
{
- if (boot_cpu_data.x86 >= 0x15)
+ if (boot_cpu_data.x86 == 0x15)
x86_pmu.get_event_constraints = amd_get_event_constraints_f15h;
return 0;
}
@@ -655,6 +727,23 @@ static int setup_perfctr_core(void)
return 0;
}
+static int setup_perfctr_nb(void)
+{
+ if (!cpu_has_perfctr_nb)
+ return -ENODEV;
+
+ x86_pmu.num_counters += AMD64_NUM_COUNTERS_NB;
+
+ if (cpu_has_perfctr_core)
+ amd_nb_event_constraint = &amd_NBPMC96;
+ else
+ amd_nb_event_constraint = &amd_NBPMC74;
+
+ printk(KERN_INFO "perf: AMD northbridge performance counters detected\n");
+
+ return 0;
+}
+
__init int amd_pmu_init(void)
{
/* Performance-monitoring supported from K7 and later: */
@@ -665,6 +754,7 @@ __init int amd_pmu_init(void)
setup_event_constraints();
setup_perfctr_core();
+ setup_perfctr_nb();
/* Events are common for all AMDs */
memcpy(hw_cache_event_ids, amd_hw_cache_event_ids,
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/