Paul suggested we allow for data addresses to be recorded along with
the traditional IPs as power can provide these.
For now, only the software pagefault events provide data addresses,
but in the future power might as well for some events.
x86 doesn't seem capable of providing this atm.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
LKML-Reference: <20090408130409.394816925@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Prepare for more generic overflow handling. The new perf_counter_overflow()
method will handle the generic bits of the counter overflow, and can return
a !0 return value, in which case the counter should be (soft) disabled, so
that it won't count until it's properly disabled.
XXX: do powerpc and swcounter
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
LKML-Reference: <20090406094517.812109629@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Implement set_perf_counter_pending() with a self-IPI so that it will
run ASAP in a usable context.
For now use a second IRQ vector, because the primary vector pokes
the apic in funny ways that seem to confuse things.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
LKML-Reference: <20090406094517.724626696@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Put in counts to tell which ips belong to what context.
-----
| | hv
| --
nr | | kernel
| --
| | user
-----
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Orig-LKML-Reference: <20090402091319.493101305@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Follow the example set by powerpc and try to play nice with oprofile
and the nmi watchdog.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Orig-LKML-Reference: <20090330171024.459968444@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Provide the x86 perf_callchain() implementation.
Code based on the ftrace/sysprof code from Soeren Sandmann Pedersen.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Cc: Soeren Sandmann Pedersen <sandmann@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <srostedt@redhat.com>
Orig-LKML-Reference: <20090330171024.341993293@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Now that Paul cleaned up the error propagation paths, pass down the
x86 error as well.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Orig-LKML-Reference: <20090330171023.792822360@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
While going over the wakeup code I noticed delayed wakeups only work
for hardware counters but basically all software counters rely on
them.
This patch unifies and generalizes the delayed wakeup to fix this
issue.
Since we're dealing with NMI context bits here, use a cmpxchg() based
single link list implementation to track counters that have pending
wakeups.
[ This should really be generic code for delayed wakeups, but since we
cannot use cmpxchg()/xchg() in generic code, I've let it live in the
perf_counter code. -- Eric Dumazet could use it to aggregate the
network wakeups. ]
Furthermore, the x86 method of using TIF flags was flawed in that its
quite possible to end up setting the bit on the idle task, loosing the
wakeup.
The powerpc method uses per-cpu storage and does appear to be
sufficient.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Paul Mackerras <paulus@samba.org>
Orig-LKML-Reference: <20090330171023.153932974@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Since the bitfields turned into a bit of a mess, remove them and rely on
good old masks.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Orig-LKML-Reference: <20090323172417.059499915@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: cleanup
Having 3 slightly different copies of the same code around does nobody
any good. First step in revamping the output format.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Orig-LKML-Reference: <20090319194233.929962222@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: modify ABI
The hardware/software classification in hw_event->type became a little
strained due to the addition of tracepoint tracing.
Instead split up the field and provide a type field to explicitly specify
the counter type, while using the event_id field to specify which event to
use.
Raw counters still work as before, only the raw config now goes into
raw_event.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Orig-LKML-Reference: <20090319194233.836807573@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: fix boot crash on Intel Perfmon Version 1 systems
Intel Perfmon v1 does not support the global MSRs, nor does
it offer the generalized MSR ranges. So support v2 and later
CPUs only.
Also mark pmc_ops as read-mostly - to avoid false cacheline
sharing.
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Fix a build warning on 32bit machines by explicitly marking the
constants as 64-bit.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
We need to ensure the enabled=0 write happens before we
start disabling the actual counters, so that a pcm_amd_enable()
will not enable one underneath us.
I think the race is impossible anyway, we always balance the
ops within any one context and perform enable() with IRQs disabled.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Use and actual unsigned long bitmap instead of casting our way around.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Jaswinder Singh Rajput <jaswinder@kernel.org>
LKML-Reference: <1236508459.22914.3645.camel@twins>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
The below completes the K7+ performance counter support:
- IRQ support
- NMI support
KernelTop output works now as well.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Jaswinder Singh Rajput <jaswinder@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <1236273633.5187.286.camel@laptop>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: cleanup
using pr_info in perf_counter.c fixes various 80 characters warnings and
also indenting for conditional statement
Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: cleanup
making decent declrations for struct pmc_x86_ops and
fix checkpatch error:
ERROR: Macros with complex values should be enclosed in parenthesis
Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: new perf_counter feature
This extends the perf_counter_hw_event struct with bits that specify
that events in user, kernel and/or hypervisor mode should not be
counted (i.e. should be excluded), and adds code to program the PMU
mode selection bits accordingly on x86 and powerpc.
For software counters, we don't currently have the infrastructure to
distinguish which mode an event occurs in, so we currently fail the
counter initialization if the setting of the hw_event.exclude_* bits
would require us to distinguish. Context switches and CPU migrations
are currently considered to occur in kernel mode.
On x86, this changes the previous policy that only root can count
kernel events. Now non-root users can count kernel events or exclude
them. Non-root users still can't use NMI events, though. On x86 we
don't appear to have any way to control whether hypervisor events are
counted or not, so hw_event.exclude_hv is ignored.
On powerpc, the selection of whether to count events in user, kernel
and/or hypervisor mode is PMU-wide, not per-counter, so this adds a
check that the hw_event.exclude_* settings are the same as other events
on the PMU. Counters being added to a group have to have the same
settings as the other hardware counters in the group. Counters and
groups can only be enabled in hw_perf_group_sched_in or power_perf_enable
if they have the same settings as any other counters already on the
PMU. If we are not running on a hypervisor, the exclude_hv setting
is ignored (by forcing it to 0) since we can't ever get any
hypervisor events.
Signed-off-by: Paul Mackerras <paulus@samba.org>
I noticed that kerneltop interrupts were accounted as NMI, but not their
perf counter origin.
Account NMI performance counter interrupts.
Signed-off-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
arch/x86/kernel/cpu/perf_counter.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
With oprofile as a module, and unloaded by profiling script,
both oprofile and kerneltop work fine.. unless you leave kerneltop
running when you start profiling, then you may see badness.
Signed-off-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Fix:
WARNING: arch/x86/kernel/built-in.o(.text+0xdd0f): Section mismatch in reference from the function pmc_generic_enable() to the function .cpuinit.text:perf_counters_lapic_init()
The function pmc_generic_enable() references
the function __cpuinit perf_counters_lapic_init().
This is often because pmc_generic_enable lacks a __cpuinit
annotation or the annotation of perf_counters_lapic_init is wrong.
Signed-off-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Ratelimit performance counter interrupts to 100KHz per CPU.
This replaces the irq-delta-time based method.
Signed-off-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Starting kerneltop with only -c 100 seems to be a bad idea, it can
easily lock the system due to perfcounter IRQ overload.
So add throttling: if a new IRQ arrives in a shorter than
PERFMON_MIN_PERIOD_NS time, turn off perfcounters and untrottle them
from the next timer tick.
Signed-off-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: extend performance counter support on x86 Intel CPUs
Modern Intel CPUs have 3 "fixed-function" performance counters, which
count these hardware events:
Instr_Retired.Any
CPU_CLK_Unhalted.Core
CPU_CLK_Unhalted.Ref
Add support for them to the performance counters subsystem.
Their use is transparent to user-space: the counter scheduler is
extended to automatically recognize the cases where a fixed-function
PMC can be utilized instead of a generic PMC. In such cases the
generic PMC is kept available for more counters.
The above fixed-function events map to these generic counter hw events:
PERF_COUNT_INSTRUCTIONS
PERF_COUNT_CPU_CYCLES
PERF_COUNT_BUS_CYCLES
(The 'bus' cycles are in reality often CPU-ish cycles, just with a fixed
frequency.)
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Generalize "bus cycles" hw events - and map them to CPU_CLK_Unhalted.Ref
on x86. (which is a good enough approximation)
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Allow lowlevel ->enable() op to return an error if a counter can not be
added. This can be used to handle counter constraints.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Enumerate fixed-mode PMCs based on CPUID, and feed that into the
perfcounter code.
Does not use fixed-mode PMCs yet.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: refactor the x86 code for fixed-mode PMCs
Extend the data structures and rename the existing facilities
to allow for a 'generic' versus 'fixed' counter distinction.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: rename include file
We'll be providing an asm/perf_counter.h to the generic perfcounter code,
so use the already existing x86 file for this purpose and rename it.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: cleanup, avoid sparse warnings, reduce kernel size a bit
Fixes these sparse warnings:
arch/x86/kernel/cpu/perf_counter.c:44:11: warning: symbol 'intel_perfmon_event_map' was not declared. Should it be static?
arch/x86/kernel/cpu/perf_counter.c:54:11: warning: symbol 'max_intel_perfmon_events' was not declared. Should it be static?
Signed-off-by: Jaswinder Singh <jaswinder@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: restructure code
Change counter math from absolute values to clear delta logic.
We try to extract elapsed deltas from the raw hw counter - and put
that into the generic counter.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: cleanup
Introduce a proper enum for the 3 states of a counter:
PERF_COUNTER_STATE_OFF = -1
PERF_COUNTER_STATE_INACTIVE = 0
PERF_COUNTER_STATE_ACTIVE = 1
and rename counter->active to counter->state and propagate the
changes everywhere.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: cleanup
Rename them to better match up the usual IRQ disable/enable APIs:
hw_perf_disable_all() => hw_perf_save_disable()
hw_perf_restore_ctrl() => hw_perf_restore()
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: add new perf-counter type
The 'CPU clock' counter counts the amount of CPU clock time that is
elapsing, in nanoseconds. (regardless of how much of it the task is
spending on a CPU executing)
This counter type is a Linux kernel based abstraction, it is available
even if the hardware does not support native hardware performance counters.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: restructure code, introduce hw_ops driver abstraction
Introduce this abstraction to handle counter details:
struct hw_perf_counter_ops {
void (*hw_perf_counter_enable) (struct perf_counter *counter);
void (*hw_perf_counter_disable) (struct perf_counter *counter);
void (*hw_perf_counter_read) (struct perf_counter *counter);
};
This will be useful to support assymetric hw details, and it will also
be useful to implement "software counters". (Counters that count kernel
managed sw events such as pagefaults, context-switches, wall-clock time
or task-local time.)
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: add group counters
This patch adds the "counter groups" abstraction.
Groups of counters behave much like normal 'single' counters, with a
few semantic and behavioral extensions on top of that.
A counter group is created by creating a new counter with the open()
syscall's group-leader group_fd file descriptor parameter pointing
to another, already existing counter.
Groups of counters are scheduled in and out in one atomic group, and
they are also roundrobin-scheduled atomically.
Counters that are member of a group can also record events with an
(atomic) extended timestamp that extends to all members of the group,
if the record type is set to PERF_RECORD_GROUP.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
