2008-02-20 06:43:25 +08:00
|
|
|
/*
|
2005-04-17 06:20:36 +08:00
|
|
|
* @file op_model_ppro.h
|
2008-08-18 20:50:31 +08:00
|
|
|
* Family 6 perfmon and architectural perfmon MSR operations
|
2005-04-17 06:20:36 +08:00
|
|
|
*
|
|
|
|
* @remark Copyright 2002 OProfile authors
|
2008-08-18 20:50:31 +08:00
|
|
|
* @remark Copyright 2008 Intel Corporation
|
2005-04-17 06:20:36 +08:00
|
|
|
* @remark Read the file COPYING
|
|
|
|
*
|
|
|
|
* @author John Levon
|
|
|
|
* @author Philippe Elie
|
|
|
|
* @author Graydon Hoare
|
2008-08-18 20:50:31 +08:00
|
|
|
* @author Andi Kleen
|
2009-05-25 21:10:32 +08:00
|
|
|
* @author Robert Richter <robert.richter@amd.com>
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/oprofile.h>
|
2008-08-18 20:50:31 +08:00
|
|
|
#include <linux/slab.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
#include <asm/ptrace.h>
|
|
|
|
#include <asm/msr.h>
|
|
|
|
#include <asm/apic.h>
|
2006-06-26 19:57:01 +08:00
|
|
|
#include <asm/nmi.h>
|
2008-02-20 06:43:25 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
#include "op_x86_model.h"
|
|
|
|
#include "op_counter.h"
|
|
|
|
|
2008-08-18 20:50:31 +08:00
|
|
|
static int num_counters = 2;
|
|
|
|
static int counter_width = 32;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2009-05-25 21:10:32 +08:00
|
|
|
#define MSR_PPRO_EVENTSEL_RESERVED ((0xFFFFFFFFULL<<32)|(1ULL<<21))
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2008-08-18 20:50:31 +08:00
|
|
|
static u64 *reset_value;
|
2008-02-20 06:43:25 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
static void ppro_fill_in_addresses(struct op_msrs * const msrs)
|
|
|
|
{
|
2006-09-26 16:52:26 +08:00
|
|
|
int i;
|
|
|
|
|
2008-08-18 20:50:31 +08:00
|
|
|
for (i = 0; i < num_counters; i++) {
|
2006-09-26 16:52:26 +08:00
|
|
|
if (reserve_perfctr_nmi(MSR_P6_PERFCTR0 + i))
|
|
|
|
msrs->counters[i].addr = MSR_P6_PERFCTR0 + i;
|
|
|
|
else
|
|
|
|
msrs->counters[i].addr = 0;
|
|
|
|
}
|
2008-02-20 06:43:25 +08:00
|
|
|
|
2008-08-18 20:50:31 +08:00
|
|
|
for (i = 0; i < num_counters; i++) {
|
2006-09-26 16:52:26 +08:00
|
|
|
if (reserve_evntsel_nmi(MSR_P6_EVNTSEL0 + i))
|
|
|
|
msrs->controls[i].addr = MSR_P6_EVNTSEL0 + i;
|
|
|
|
else
|
|
|
|
msrs->controls[i].addr = 0;
|
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2009-05-26 01:31:44 +08:00
|
|
|
static void ppro_setup_ctrs(struct op_x86_model_spec const *model,
|
|
|
|
struct op_msrs const * const msrs)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2009-05-25 21:10:32 +08:00
|
|
|
u64 val;
|
2005-04-17 06:20:36 +08:00
|
|
|
int i;
|
|
|
|
|
2008-08-18 20:50:31 +08:00
|
|
|
if (!reset_value) {
|
2008-11-10 16:05:37 +08:00
|
|
|
reset_value = kmalloc(sizeof(reset_value[0]) * num_counters,
|
2008-08-18 20:50:31 +08:00
|
|
|
GFP_ATOMIC);
|
|
|
|
if (!reset_value)
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (cpu_has_arch_perfmon) {
|
|
|
|
union cpuid10_eax eax;
|
|
|
|
eax.full = cpuid_eax(0xa);
|
2009-02-20 00:34:03 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* For Core2 (family 6, model 15), don't reset the
|
|
|
|
* counter width:
|
|
|
|
*/
|
|
|
|
if (!(eax.split.version_id == 0 &&
|
|
|
|
current_cpu_data.x86 == 6 &&
|
|
|
|
current_cpu_data.x86_model == 15)) {
|
|
|
|
|
|
|
|
if (counter_width < eax.split.bit_width)
|
|
|
|
counter_width = eax.split.bit_width;
|
|
|
|
}
|
2008-08-18 20:50:31 +08:00
|
|
|
}
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/* clear all counters */
|
2009-07-08 01:25:39 +08:00
|
|
|
for (i = 0; i < num_counters; ++i) {
|
2009-06-04 08:36:44 +08:00
|
|
|
if (unlikely(!msrs->controls[i].addr))
|
2006-09-26 16:52:26 +08:00
|
|
|
continue;
|
2009-05-25 21:10:32 +08:00
|
|
|
rdmsrl(msrs->controls[i].addr, val);
|
|
|
|
val &= model->reserved;
|
|
|
|
wrmsrl(msrs->controls[i].addr, val);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2008-02-20 06:43:25 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/* avoid a false detection of ctr overflows in NMI handler */
|
2008-08-18 20:50:31 +08:00
|
|
|
for (i = 0; i < num_counters; ++i) {
|
2009-06-04 08:36:44 +08:00
|
|
|
if (unlikely(!msrs->counters[i].addr))
|
2006-09-26 16:52:26 +08:00
|
|
|
continue;
|
2008-08-18 20:50:31 +08:00
|
|
|
wrmsrl(msrs->counters[i].addr, -1LL);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/* enable active counters */
|
2008-08-18 20:50:31 +08:00
|
|
|
for (i = 0; i < num_counters; ++i) {
|
2009-06-04 08:36:44 +08:00
|
|
|
if (counter_config[i].enabled && msrs->counters[i].addr) {
|
2005-04-17 06:20:36 +08:00
|
|
|
reset_value[i] = counter_config[i].count;
|
2008-08-18 20:50:31 +08:00
|
|
|
wrmsrl(msrs->counters[i].addr, -reset_value[i]);
|
2009-05-25 21:10:32 +08:00
|
|
|
rdmsrl(msrs->controls[i].addr, val);
|
|
|
|
val &= model->reserved;
|
|
|
|
val |= op_x86_get_ctrl(model, &counter_config[i]);
|
|
|
|
wrmsrl(msrs->controls[i].addr, val);
|
2006-09-26 16:52:26 +08:00
|
|
|
} else {
|
|
|
|
reset_value[i] = 0;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2008-02-20 06:43:25 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
static int ppro_check_ctrs(struct pt_regs * const regs,
|
|
|
|
struct op_msrs const * const msrs)
|
|
|
|
{
|
2008-11-07 21:02:49 +08:00
|
|
|
u64 val;
|
2005-04-17 06:20:36 +08:00
|
|
|
int i;
|
2008-02-20 06:43:25 +08:00
|
|
|
|
perfcounters: fix "perf counters kills oprofile" bug, v2
Impact: fix kernel crash
Both oprofile and perfcounters register an NMI die handler, but only one
can handle the NMI. Conveniently, oprofile unregisters it's notifier
when not actively in use, so setting it's notifier priority higher than
perfcounter's allows oprofile to borrow the NMI for the duration of it's
run. Tested/works both as module and built-in.
While testing, I found that if kerneltop was generating NMIs at very
high frequency, the kernel may panic when oprofile registered it's
handler. This turned out to be because oprofile registers it's handler
before reset_value has been allocated, so if an NMI comes in while it's
still setting up, kabOom. Rather than try more invasive changes, I
followed the lead of other places in op_model_ppro.c, and simply
returned in that highly unlikely event. (debug warnings attached)
Signed-off-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-05 22:23:08 +08:00
|
|
|
/*
|
|
|
|
* This can happen if perf counters are in use when
|
|
|
|
* we steal the die notifier NMI.
|
|
|
|
*/
|
|
|
|
if (unlikely(!reset_value))
|
|
|
|
goto out;
|
|
|
|
|
2009-07-08 01:25:39 +08:00
|
|
|
for (i = 0; i < num_counters; ++i) {
|
2006-09-26 16:52:26 +08:00
|
|
|
if (!reset_value[i])
|
|
|
|
continue;
|
2008-11-07 21:02:49 +08:00
|
|
|
rdmsrl(msrs->counters[i].addr, val);
|
2009-05-25 23:59:06 +08:00
|
|
|
if (val & (1ULL << (counter_width - 1)))
|
|
|
|
continue;
|
|
|
|
oprofile_add_sample(regs, i);
|
|
|
|
wrmsrl(msrs->counters[i].addr, -reset_value[i]);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
perfcounters: fix "perf counters kills oprofile" bug, v2
Impact: fix kernel crash
Both oprofile and perfcounters register an NMI die handler, but only one
can handle the NMI. Conveniently, oprofile unregisters it's notifier
when not actively in use, so setting it's notifier priority higher than
perfcounter's allows oprofile to borrow the NMI for the duration of it's
run. Tested/works both as module and built-in.
While testing, I found that if kerneltop was generating NMIs at very
high frequency, the kernel may panic when oprofile registered it's
handler. This turned out to be because oprofile registers it's handler
before reset_value has been allocated, so if an NMI comes in while it's
still setting up, kabOom. Rather than try more invasive changes, I
followed the lead of other places in op_model_ppro.c, and simply
returned in that highly unlikely event. (debug warnings attached)
Signed-off-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-05 22:23:08 +08:00
|
|
|
out:
|
2005-04-17 06:20:36 +08:00
|
|
|
/* Only P6 based Pentium M need to re-unmask the apic vector but it
|
|
|
|
* doesn't hurt other P6 variant */
|
|
|
|
apic_write(APIC_LVTPC, apic_read(APIC_LVTPC) & ~APIC_LVT_MASKED);
|
|
|
|
|
|
|
|
/* We can't work out if we really handled an interrupt. We
|
|
|
|
* might have caught a *second* counter just after overflowing
|
|
|
|
* the interrupt for this counter then arrives
|
|
|
|
* and we don't find a counter that's overflowed, so we
|
|
|
|
* would return 0 and get dazed + confused. Instead we always
|
|
|
|
* assume we found an overflow. This sucks.
|
|
|
|
*/
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2008-02-20 06:43:25 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
static void ppro_start(struct op_msrs const * const msrs)
|
|
|
|
{
|
2009-05-26 00:11:52 +08:00
|
|
|
u64 val;
|
2006-09-29 17:00:01 +08:00
|
|
|
int i;
|
2006-09-26 16:52:26 +08:00
|
|
|
|
2008-12-02 14:21:21 +08:00
|
|
|
if (!reset_value)
|
|
|
|
return;
|
2008-08-18 20:50:31 +08:00
|
|
|
for (i = 0; i < num_counters; ++i) {
|
2006-09-29 17:00:01 +08:00
|
|
|
if (reset_value[i]) {
|
2009-05-26 00:11:52 +08:00
|
|
|
rdmsrl(msrs->controls[i].addr, val);
|
|
|
|
val |= ARCH_PERFMON_EVENTSEL0_ENABLE;
|
|
|
|
wrmsrl(msrs->controls[i].addr, val);
|
2006-09-29 17:00:01 +08:00
|
|
|
}
|
2006-09-26 16:52:26 +08:00
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static void ppro_stop(struct op_msrs const * const msrs)
|
|
|
|
{
|
2009-05-26 00:11:52 +08:00
|
|
|
u64 val;
|
2006-09-29 17:00:01 +08:00
|
|
|
int i;
|
2006-09-26 16:52:26 +08:00
|
|
|
|
2008-12-02 14:21:21 +08:00
|
|
|
if (!reset_value)
|
|
|
|
return;
|
2008-08-18 20:50:31 +08:00
|
|
|
for (i = 0; i < num_counters; ++i) {
|
2006-09-29 17:00:01 +08:00
|
|
|
if (!reset_value[i])
|
|
|
|
continue;
|
2009-05-26 00:11:52 +08:00
|
|
|
rdmsrl(msrs->controls[i].addr, val);
|
|
|
|
val &= ~ARCH_PERFMON_EVENTSEL0_ENABLE;
|
|
|
|
wrmsrl(msrs->controls[i].addr, val);
|
2006-09-26 16:52:26 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void ppro_shutdown(struct op_msrs const * const msrs)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
2009-07-08 01:25:39 +08:00
|
|
|
for (i = 0; i < num_counters; ++i) {
|
2009-06-04 08:36:44 +08:00
|
|
|
if (msrs->counters[i].addr)
|
2006-09-26 16:52:26 +08:00
|
|
|
release_perfctr_nmi(MSR_P6_PERFCTR0 + i);
|
|
|
|
}
|
2009-07-08 01:25:39 +08:00
|
|
|
for (i = 0; i < num_counters; ++i) {
|
2009-06-04 08:36:44 +08:00
|
|
|
if (msrs->controls[i].addr)
|
2006-09-26 16:52:26 +08:00
|
|
|
release_evntsel_nmi(MSR_P6_EVNTSEL0 + i);
|
|
|
|
}
|
2008-08-18 20:50:31 +08:00
|
|
|
if (reset_value) {
|
|
|
|
kfree(reset_value);
|
|
|
|
reset_value = NULL;
|
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
2009-07-09 21:12:35 +08:00
|
|
|
struct op_x86_model_spec op_ppro_spec = {
|
2009-05-14 23:10:52 +08:00
|
|
|
.num_counters = 2,
|
|
|
|
.num_controls = 2,
|
2009-05-25 21:10:32 +08:00
|
|
|
.reserved = MSR_PPRO_EVENTSEL_RESERVED,
|
2008-10-16 04:19:41 +08:00
|
|
|
.fill_in_addresses = &ppro_fill_in_addresses,
|
|
|
|
.setup_ctrs = &ppro_setup_ctrs,
|
|
|
|
.check_ctrs = &ppro_check_ctrs,
|
|
|
|
.start = &ppro_start,
|
|
|
|
.stop = &ppro_stop,
|
|
|
|
.shutdown = &ppro_shutdown
|
2008-08-18 20:50:31 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Architectural performance monitoring.
|
|
|
|
*
|
|
|
|
* Newer Intel CPUs (Core1+) have support for architectural
|
|
|
|
* events described in CPUID 0xA. See the IA32 SDM Vol3b.18 for details.
|
|
|
|
* The advantage of this is that it can be done without knowing about
|
|
|
|
* the specific CPU.
|
|
|
|
*/
|
|
|
|
|
2008-10-13 03:12:34 +08:00
|
|
|
static void arch_perfmon_setup_counters(void)
|
2008-08-18 20:50:31 +08:00
|
|
|
{
|
|
|
|
union cpuid10_eax eax;
|
|
|
|
|
|
|
|
eax.full = cpuid_eax(0xa);
|
|
|
|
|
|
|
|
/* Workaround for BIOS bugs in 6/15. Taken from perfmon2 */
|
|
|
|
if (eax.split.version_id == 0 && current_cpu_data.x86 == 6 &&
|
|
|
|
current_cpu_data.x86_model == 15) {
|
|
|
|
eax.split.version_id = 2;
|
|
|
|
eax.split.num_counters = 2;
|
|
|
|
eax.split.bit_width = 40;
|
|
|
|
}
|
|
|
|
|
|
|
|
num_counters = eax.split.num_counters;
|
|
|
|
|
|
|
|
op_arch_perfmon_spec.num_counters = num_counters;
|
|
|
|
op_arch_perfmon_spec.num_controls = num_counters;
|
|
|
|
}
|
|
|
|
|
2008-10-13 03:12:34 +08:00
|
|
|
static int arch_perfmon_init(struct oprofile_operations *ignore)
|
|
|
|
{
|
|
|
|
arch_perfmon_setup_counters();
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-08-18 20:50:31 +08:00
|
|
|
struct op_x86_model_spec op_arch_perfmon_spec = {
|
2009-05-25 21:10:32 +08:00
|
|
|
.reserved = MSR_PPRO_EVENTSEL_RESERVED,
|
2008-10-13 03:12:34 +08:00
|
|
|
.init = &arch_perfmon_init,
|
2008-08-18 20:50:31 +08:00
|
|
|
/* num_counters/num_controls filled in at runtime */
|
2008-09-05 23:12:36 +08:00
|
|
|
.fill_in_addresses = &ppro_fill_in_addresses,
|
2008-08-18 20:50:31 +08:00
|
|
|
/* user space does the cpuid check for available events */
|
2008-09-05 23:12:36 +08:00
|
|
|
.setup_ctrs = &ppro_setup_ctrs,
|
|
|
|
.check_ctrs = &ppro_check_ctrs,
|
|
|
|
.start = &ppro_start,
|
|
|
|
.stop = &ppro_stop,
|
|
|
|
.shutdown = &ppro_shutdown
|
2005-04-17 06:20:36 +08:00
|
|
|
};
|