path: root/man2/perf_event_open.2

.\" Copyright (c) 2012, Vincent Weaver
.\"
.\" %%%LICENSE_START(GPLv2+_DOC_FULL)
.\" This is free documentation; you can redistribute it and/or
.\" modify it under the terms of the GNU General Public License as
.\" published by the Free Software Foundation; either version 2 of
.\" the License, or (at your option) any later version.
.\"
.\" The GNU General Public License's references to "object code"
.\" and "executables" are to be interpreted as the output of any
.\" document formatting or typesetting system, including
.\" intermediate and printed output.
.\"
.\" This manual is distributed in the hope that it will be useful,
.\" but WITHOUT ANY WARRANTY; without even the implied warranty of
.\" MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
.\" GNU General Public License for more details.
.\"
.\" You should have received a copy of the GNU General Public
.\" License along with this manual; if not, see
.\" <http://www.gnu.org/licenses/>.
.\" %%%LICENSE_END
.\"
.\" This document is based on the perf_event.h header file, the
.\" tools/perf/design.txt file, and a lot of bitter experience.
.\"
.TH PERF_EVENT_OPEN 2 2013-07-16 "Linux" "Linux Programmer's Manual"
.SH NAME
perf_event_open \- set up performance monitoring
.SH SYNOPSIS
.nf
.B #include <linux/perf_event.h>
.B #include <linux/hw_breakpoint.h>
.sp
.BI "int perf_event_open(struct perf_event_attr *" attr ,
.BI "                    pid_t " pid ", int " cpu ", int " group_fd ,
.BI "                    unsigned long " flags  );
.fi

.IR Note :
There is no glibc wrapper for this system call; see NOTES.
.SH DESCRIPTION
Given a list of parameters,
.BR perf_event_open ()
returns a file descriptor, for use in subsequent system calls
.RB ( read "(2), " mmap "(2), " prctl "(2), " fcntl "(2), etc.)."
.PP
A call to
.BR perf_event_open ()
creates a file descriptor that allows measuring performance
information.
Each file descriptor corresponds to one
event that is measured; these can be grouped together
to measure multiple events simultaneously.
.PP
Events can be enabled and disabled in two ways: via
.BR ioctl (2)
and via
.BR prctl (2) .
When an event is disabled it does not count or generate overflows but does
continue to exist and maintain its count value.
.PP
Events come in two flavors: counting and sampled.
A
.I counting
event is one that is used for counting the aggregate number of events
that occur.
In general, counting event results are gathered with a
.BR read (2)
call.
A
.I sampling
event periodically writes measurements to a buffer that can then
be accessed via
.BR  mmap (2) .
.SS Arguments
.P
The argument
.I pid
allows events to be attached to processes in various ways.
If
.I pid
is 0, measurements happen on the current thread, if
.I pid
is greater than 0, the process indicated by
.I pid
is measured, and if
.I pid
is \-1, all processes are counted.

The
.I cpu
argument allows measurements to be specific to a CPU.
If
.I cpu
is greater than or equal to 0,
measurements are restricted to the specified CPU;
if
.I cpu
is \-1, the events are measured on all CPUs.
.P
Note that the combination of
.IR pid " == \-1"
and
.IR cpu " == \-1"
is not valid.
.P
A
.IR pid " > 0"
and
.IR cpu " == \-1"
setting measures per-process and follows that process to whatever CPU the
process gets scheduled to.
Per-process events can be created by any user.
.P
A
.IR pid " == \-1"
and
.IR cpu " >= 0"
setting is per-CPU and measures all processes on the specified CPU.
Per-CPU events need the
.B CAP_SYS_ADMIN
capability or a
.I /proc/sys/kernel/perf_event_paranoid
value of less than 1.
.P
The
.I group_fd
argument allows event groups to be created.
An event group has one event which is the group leader.
The leader is created first, with
.IR group_fd " = \-1."
The rest of the group members are created with subsequent
.BR perf_event_open ()
calls with
.IR group_fd
being set to the fd of the group leader.
(A single event on its own is created with
.IR group_fd " = \-1"
and is considered to be a group with only 1 member.)
An event group is scheduled onto the CPU as a unit: it will
be put onto the CPU only if all of the events in the group can be put onto
the CPU.
This means that the values of the member events can be
meaningfully compared, added, divided (to get ratios), etc., with each
other, since they have counted events for the same set of executed
instructions.
.P
The
.I flags
argument is formed by ORing together zero or more of the following values:
.TP
.BR PERF_FLAG_FD_NO_GROUP
.\" FIXME The following sentence is unclear
This flag allows creating an event as part of an event group but
having no group leader.
It is unclear why this is useful.
.\" FIXME So, why is it useful?
.TP
.BR PERF_FLAG_FD_OUTPUT
This flag re-routes the output from an event to the group leader.
.TP
.BR PERF_FLAG_PID_CGROUP " (Since Linux 2.6.39)."
This flag activates per-container system-wide monitoring.
A container
is an abstraction that isolates a set of resources for finer grain
control (CPUs, memory, etc.).
In this mode, the event is measured
only if the thread running on the monitored CPU belongs to the designated
container (cgroup).
The cgroup is identified by passing a file descriptor
opened on its directory in the cgroupfs filesystem.
For instance, if the
cgroup to monitor is called
.IR test ,
then a file descriptor opened on
.I /dev/cgroup/test
(assuming cgroupfs is mounted on
.IR /dev/cgroup )
must be passed as the
.I pid
parameter.
cgroup monitoring is available only
for system-wide events and may therefore require extra permissions.
.P
The
.I perf_event_attr
structure provides detailed configuration information
for the event being created.

.in +4n
.nf
struct perf_event_attr {
    __u32     type;         /* Type of event */
    __u32     size;         /* Size of attribute structure */
    __u64     config;       /* Type-specific configuration */

    union {
        __u64 sample_period;    /* Period of sampling */
        __u64 sample_freq;      /* Frequency of sampling */
    };

    __u64     sample_type;  /* Specifies values included in sample */
    __u64     read_format;  /* Specifies values returned in read */

    __u64     disabled       : 1,   /* off by default */
              inherit        : 1,   /* children inherit it */
              pinned         : 1,   /* must always be on PMU */
              exclusive      : 1,   /* only group on PMU */
              exclude_user   : 1,   /* don't count user */
              exclude_kernel : 1,   /* don't count kernel */
              exclude_hv     : 1,   /* don't count hypervisor */
              exclude_idle   : 1,   /* don't count when idle */
              mmap           : 1,   /* include mmap data */
              comm           : 1,   /* include comm data */
              freq           : 1,   /* use freq, not period */
              inherit_stat   : 1,   /* per task counts */
              enable_on_exec : 1,   /* next exec enables */
              task           : 1,   /* trace fork/exit */
              watermark      : 1,   /* wakeup_watermark */
              precise_ip     : 2,   /* skid constraint */
              mmap_data      : 1,   /* non-exec mmap data */
              sample_id_all  : 1,   /* sample_type all events */
              exclude_host   : 1,   /* don't count in host */
              exclude_guest  : 1,   /* don't count in guest */
              exclude_callchain_kernel : 1,
                                    /* exclude kernel callchains */
              exclude_callchain_user   : 1,
	                            /* exclude user callchains */
              __reserved_1   : 41;

    union {
        __u32 wakeup_events;    /* wakeup every n events */
        __u32 wakeup_watermark; /* bytes before wakeup */
    };

    __u32     bp_type;          /* breakpoint type */

    union {
        __u64 bp_addr;          /* breakpoint address */
        __u64 config1;          /* extension of config */
    };

    union {
        __u64 bp_len;           /* breakpoint length */
        __u64 config2;          /* extension of config1 */
    };
    __u64   branch_sample_type; /* enum perf_branch_sample_type */
    __u64   sample_regs_user;   /* user regs to dump on samples */
    __u32   sample_stack_user;  /* size of stack to dump on
                                   samples */
    __u32   __reserved_2;       /* Align to u64 */

};
.fi
.in

The fields of the
.I perf_event_attr
structure are described in more detail below:
.TP
.I type
This field specifies the overall event type.
It has one of the following values:
.RS
.TP
.B PERF_TYPE_HARDWARE
This indicates one of the "generalized" hardware events provided
by the kernel.
See the
.I config
field definition for more details.
.TP
.B PERF_TYPE_SOFTWARE
This indicates one of the software-defined events provided by the kernel
(even if no hardware support is available).
.TP
.B PERF_TYPE_TRACEPOINT
This indicates a tracepoint
provided by the kernel tracepoint infrastructure.
.TP
.B PERF_TYPE_HW_CACHE
This indicates a hardware cache event.
This has a special encoding, described in the
.I config
field definition.
.TP
.B PERF_TYPE_RAW
This indicates a "raw" implementation-specific event in the
.IR config " field."
.TP
.BR PERF_TYPE_BREAKPOINT " (Since Linux 2.6.33)"
This indicates a hardware breakpoint as provided by the CPU.
Breakpoints can be read/write accesses to an address as well as
execution of an instruction address.
.TP
.RB "dynamic PMU"
Since Linux 2.6.39,
.BR perf_event_open ()
can support multiple PMUs.
To enable this, a value exported by the kernel can be used in the
.I type
field to indicate which PMU to use.
The value to use can be found in the sysfs filesystem:
there is a subdirectory per PMU instance under
.IR /sys/bus/event_source/devices .
In each sub-directory there is a
.I type
file whose content is an integer that can be used in the
.I type
field.
For instance,
.I /sys/bus/event_source/devices/cpu/type
contains the value for the core CPU PMU, which is usually 4.
.RE
.TP
.I "size"
The size of the
.I perf_event_attr
structure for forward/backward compatibility.
Set this using
.I sizeof(struct perf_event_attr)
to allow the kernel to see
the struct size at the time of compilation.

The related define
.B PERF_ATTR_SIZE_VER0
is set to 64; this was the size of the first published struct.
.B PERF_ATTR_SIZE_VER1
is 72, corresponding to the addition of breakpoints in Linux 2.6.33.
.B PERF_ATTR_SIZE_VER2
is 80 corresponding to the addition of branch sampling in Linux 3.4.
.B PERF_ATR_SIZE_VER3
is 96 corresponding to the addition
of
.I sample_regs_user
and
.I sample_stack_user
in Linux 3.7.
.TP
.I "config"
This specifies which event you want, in conjunction with
the
.I type
field.
The
.IR config1 " and " config2
fields are also taken into account in cases where 64 bits is not
enough to fully specify the event.
The encoding of these fields are event dependent.

The most significant bit (bit 63) of
.I config
signifies CPU-specific (raw) counter configuration data;
if the most significant bit is unset, the next 7 bits are an event
type and the rest of the bits are the event identifier.

There are various ways to set the
.I config
field that are dependent on the value of the previously
described
.I type
field.
What follows are various possible settings for
.I config
separated out by
.IR type .

If
.I type
is
.BR PERF_TYPE_HARDWARE ,
we are measuring one of the generalized hardware CPU events.
Not all of these are available on all platforms.
Set
.I config
to one of the following:
.RS 12
.TP
.B PERF_COUNT_HW_CPU_CYCLES
Total cycles.
Be wary of what happens during CPU frequency scaling
.TP
.B PERF_COUNT_HW_INSTRUCTIONS
Retired instructions.
Be careful, these can be affected by various
issues, most notably hardware interrupt counts
.TP
.B PERF_COUNT_HW_CACHE_REFERENCES
Cache accesses.
Usually this indicates Last Level Cache accesses but this may
vary depending on your CPU.
This may include prefetches and coherency messages; again this
depends on the design of your CPU.
.TP
.B PERF_COUNT_HW_CACHE_MISSES
Cache misses.
Usually this indicates Last Level Cache misses; this is intended to be
used in conjunction with the
.B PERF_COUNT_HW_CACHE_REFERENCES
event to calculate cache miss rates.
.TP
.B PERF_COUNT_HW_BRANCH_INSTRUCTIONS
Retired branch instructions.
Prior to Linux 2.6.34, this used
the wrong event on AMD processors.
.TP
.B PERF_COUNT_HW_BRANCH_MISSES
Mispredicted branch instructions.
.TP
.B PERF_COUNT_HW_BUS_CYCLES
Bus cycles, which can be different from total cycles.
.TP
.BR PERF_COUNT_HW_STALLED_CYCLES_FRONTEND " (Since Linux 3.0)"
Stalled cycles during issue.
.TP
.BR PERF_COUNT_HW_STALLED_CYCLES_BACKEND  " (Since Linux 3.0)"
Stalled cycles during retirement.
.TP
.BR PERF_COUNT_HW_REF_CPU_CYCLES  " (Since Linux 3.3)"
Total cycles; not affected by CPU frequency scaling.
.RE
.IP
If
.I type
is
.BR PERF_TYPE_SOFTWARE ,
we are measuring software events provided by the kernel.
Set
.I config
to one of the following:
.RS 12
.TP
.B PERF_COUNT_SW_CPU_CLOCK
This reports the CPU clock, a high-resolution per-CPU timer.
.TP
.B PERF_COUNT_SW_TASK_CLOCK
This reports a clock count specific to the task that is running.
.TP
.B PERF_COUNT_SW_PAGE_FAULTS
This reports the number of page faults.
.TP
.B PERF_COUNT_SW_CONTEXT_SWITCHES
This counts context switches.
Until Linux 2.6.34, these were all reported as user-space
events, after that they are reported as happening in the kernel.
.TP
.B PERF_COUNT_SW_CPU_MIGRATIONS
This reports the number of times the process
has migrated to a new CPU.
.TP
.B PERF_COUNT_SW_PAGE_FAULTS_MIN
This counts the number of minor page faults.
These did not require disk I/O to handle.
.TP
.B PERF_COUNT_SW_PAGE_FAULTS_MAJ
This counts the number of major page faults.
These required disk I/O to handle.
.TP
.BR PERF_COUNT_SW_ALIGNMENT_FAULTS " (Since Linux 2.6.33)"
This counts the number of alignment faults.
These happen when unaligned memory accesses happen; the kernel
can handle these but it reduces performance.
This happens only on some architectures (never on x86).
.TP
.BR PERF_COUNT_SW_EMULATION_FAULTS " (Since Linux 2.6.33)"
This counts the number of emulation faults.
The kernel sometimes traps on unimplemented instructions
and emulates them for user space.
This can negatively impact performance.
.RE

.RS
If
.I type
is
.BR PERF_TYPE_TRACEPOINT ,
then we are measuring kernel tracepoints.
The value to use in
.I config
can be obtained from under debugfs
.I tracing/events/*/*/id
if ftrace is enabled in the kernel.
.RE

.RS
If
.I type
is
.BR PERF_TYPE_HW_CACHE ,
then we are measuring a hardware CPU cache event.
To calculate the appropriate
.I config
value use the following equation:
.RS 4
.nf

    (perf_hw_cache_id) | (perf_hw_cache_op_id << 8) |
    (perf_hw_cache_op_result_id << 16)
.fi
.P
where
.I perf_hw_cache_id
is one of:
.RS 4
.TP
.B PERF_COUNT_HW_CACHE_L1D
for measuring Level 1 Data Cache
.TP
.B PERF_COUNT_HW_CACHE_L1I
for measuring Level 1 Instruction Cache
.TP
.B PERF_COUNT_HW_CACHE_LL
for measuring Last-Level Cache
.TP
.B PERF_COUNT_HW_CACHE_DTLB
for measuring the Data TLB
.TP
.B PERF_COUNT_HW_CACHE_ITLB
for measuring the Instruction TLB
.TP
.B PERF_COUNT_HW_CACHE_BPU
for measuring the branch prediction unit
.TP
.BR PERF_COUNT_HW_CACHE_NODE " (Since Linux 3.0)"
for measuring local memory accesses
.RE
.P
and
.I perf_hw_cache_op_id
is one of
.RS 4
.TP
.B PERF_COUNT_HW_CACHE_OP_READ
for read accesses
.TP
.B PERF_COUNT_HW_CACHE_OP_WRITE
for write accesses
.TP
.B PERF_COUNT_HW_CACHE_OP_PREFETCH
for prefetch accesses
.RE
.P
and
.I perf_hw_cache_op_result_id
is one of
.RS 4
.TP
.B PERF_COUNT_HW_CACHE_RESULT_ACCESS
to measure accesses
.TP
.B PERF_COUNT_HW_CACHE_RESULT_MISS
to measure misses
.RE
.RE

If
.I type
is
.BR PERF_TYPE_RAW ,
then a custom "raw"
.I config
value is needed.
Most CPUs support events that are not covered by the "generalized" events.
These are implementation defined; see your CPU manual (for example
the Intel Volume 3B documentation or the AMD BIOS and Kernel Developer
Guide).
The libpfm4 library can be used to translate from the name in the
architectural manuals to the raw hex value
.BR perf_event_open ()
expects in this field.

If
.I type
is
.BR PERF_TYPE_BREAKPOINT ,
then leave
.I config
set to zero.
Its parameters are set in other places.
.RE
.TP
.IR sample_period ", " sample_freq
A "sampling" counter is one that generates an interrupt
every N events, where N is given by
.IR sample_period .
A sampling counter has
.IR sample_period " > 0."
When an overflow interrupt occurs, requested data is recorded
in the mmap buffer.
The
.I sample_type
field controls what data is recorded on each interrupt.

.I sample_freq
can be used if you wish to use frequency rather than period.
In this case you set the
.I freq
flag.
The kernel will adjust the sampling period
to try and achieve the desired rate.
The rate of adjustment is a
timer tick.
.TP
.I "sample_type"
The various bits in this field specify which values to include
in the sample.
They will be recorded in a ring-buffer,
which is available to user space using
.BR mmap (2).
The order in which the values are saved in the
sample are documented in the MMAP Layout subsection below;
it is not the
.I "enum perf_event_sample_format"
order.
.RS
.TP
.B PERF_SAMPLE_IP
Records instruction pointer.
.TP
.B PERF_SAMPLE_TID
Records the process and thread IDs.
.TP
.B PERF_SAMPLE_TIME
Records a timestamp.
.TP
.B PERF_SAMPLE_ADDR
Records an address, if applicable.
.TP
.B PERF_SAMPLE_READ
Record counter values for all events in a group, not just the group leader.
.TP
.B PERF_SAMPLE_CALLCHAIN
Records the callchain (stack backtrace).
.TP
.B PERF_SAMPLE_ID
Records a unique ID for the opened event's group leader.
.TP
.B PERF_SAMPLE_CPU
Records CPU number.
.TP
.B PERF_SAMPLE_PERIOD
Records the current sampling period.
.TP
.B PERF_SAMPLE_STREAM_ID
Records a unique ID for the opened event.
Unlike
.B PERF_SAMPLE_ID
the actual ID is returned, not the group leader.
This ID is the same as the one returned by PERF_FORMAT_ID.
.TP
.B PERF_SAMPLE_RAW
Records additional data, if applicable.
Usually returned by tracepoint events.
.TP
.BR PERF_SAMPLE_BRANCH_STACK " (Since Linux 3.4)"
This provides a record of recent branches, as provided
by CPU branch sampling hardware (such as Intel Last Branch Record).
Not all hardware supports this feature.

See the
.I branch_sample_type
field for how to filter which branches are reported.
.TP
.BR PERF_SAMPLE_REGS_USER " (Since Linux 3.7)"
Records the current user-level CPU register state
(the values in the process before the kernel was called).
.TP
.BR PERF_SAMPLE_STACK_USER " (Since Linux 3.7)"
Records the user level stack, allowing stack unwinding.
.TP
.BR PERF_SAMPLE_WEIGHT " (Since Linux 3.10)"
Records a hardware provided weight value that expresses how
costly the sampled event was.
This allows the hardware to highlight expensive events in
a profile.
.TP
.BR PERF_SAMPLE_DATA_SRC " (Since Linux 3.10)"
Records the data source: where in the memory hierarchy
the data associated with the sampled instruction came from.
This is only available if the underlying hardware
supports this feature.
.RE
.TP
.IR "read_format"
This field specifies the format of the data returned by
.BR read (2)
on a
.BR perf_event_open ()
file descriptor.
.RS
.TP
.B PERF_FORMAT_TOTAL_TIME_ENABLED
Adds the 64-bit
.I time_enabled
field.
This can be used to calculate estimated totals if
the PMU is overcommitted and multiplexing is happening.
.TP
.B PERF_FORMAT_TOTAL_TIME_RUNNING
Adds the 64-bit
.I time_running
field.
This can be used to calculate estimated totals if
the PMU is overcommitted and  multiplexing is happening.
.TP
.B PERF_FORMAT_ID
Adds a 64-bit unique value that corresponds to the event group.
.TP
.B PERF_FORMAT_GROUP
Allows all counter values in an event group to be read with one read.
.RE
.TP
.IR "disabled"
The
.I disabled
bit specifies whether the counter starts out disabled or enabled.
If disabled, the event can later be enabled by
.BR ioctl (2),
.BR prctl (2),
or
.IR enable_on_exec .
.TP
.IR "inherit"
The
.I inherit
bit specifies that this counter should count events of child
tasks as well as the task specified.
This applies only to new children, not to any existing children at
the time the counter is created (nor to any new children of
existing children).

Inherit does not work for some combinations of
.IR read_format s,
such as
.BR PERF_FORMAT_GROUP .
.TP
.IR "pinned"
The
.I pinned
bit specifies that the counter should always be on the CPU if at all
possible.
It applies only to hardware counters and only to group leaders.
If a pinned counter cannot be put onto the CPU (e.g., because there are
not enough hardware counters or because of a conflict with some other
event), then the counter goes into an 'error' state, where reads
return end-of-file (i.e.,
.BR read (2)
returns 0) until the counter is subsequently enabled or disabled.
.TP
.IR "exclusive"
The
.I exclusive
bit specifies that when this counter's group is on the CPU,
it should be the only group using the CPU's counters.
In the future this may allow monitoring programs to
support PMU features that need to run alone so that they do not
disrupt other hardware counters.
.TP
.IR "exclude_user"
If this bit is set, the count excludes events that happen in user space.
.TP
.IR "exclude_kernel"
If this bit is set, the count excludes events that happen in kernel-space.
.TP
.IR "exclude_hv"
If this bit is set, the count excludes events that happen in the
hypervisor.
This is mainly for PMUs that have built-in support for handling this
(such as POWER).
Extra support is needed for handling hypervisor measurements on most
machines.
.TP
.IR "exclude_idle"
If set, don't count when the CPU is idle.
.TP
.IR "mmap"
The
.I mmap
bit enables recording of exec mmap events.
.TP
.IR "comm"
The
.I comm
bit enables tracking of process command name as modified by the
.IR exec (2)
and
.IR prctl (PR_SET_NAME)
system calls.
Unfortunately for tools,
there is no way to distinguish one system call versus the other.
.TP
.IR "freq"
If this bit is set, then
.I sample_frequency
not
.I sample_period
is used when setting up the sampling interval.
.TP
.IR "inherit_stat"
This bit enables saving of event counts on context switch for
inherited tasks.
This is meaningful only if the
.I inherit
field is set.
.TP
.IR "enable_on_exec"
If this bit is set, a counter is automatically
enabled after a call to
.BR exec (2).
.TP
.IR "task"
If this bit is set, then
fork/exit notifications are included in the ring buffer.
.TP
.IR "watermark"
If set, have a sampling interrupt happen when we cross the
.I wakeup_watermark
boundary.
Otherwise interrupts happen after
.I wakeup_events
samples.
.TP
.IR "precise_ip" " (Since Linux 2.6.35)"
This controls the amount of skid.
Skid is how many instructions
execute between an event of interest happening and the kernel
being able to stop and record the event.
Smaller skid is
better and allows more accurate reporting of which events
correspond to which instructions, but hardware is often limited
with how small this can be.

The values of this are the following:
.RS
.TP
0 -
.B SAMPLE_IP
can have arbitrary skid
.TP
1 -
.B SAMPLE_IP
must have constant skid
.TP
2 -
.B SAMPLE_IP
requested to have 0 skid
.TP
3 -
.B SAMPLE_IP
must have 0 skid.
See also
.BR PERF_RECORD_MISC_EXACT_IP .
.RE
.TP
.IR "mmap_data" " (Since Linux 2.6.36)"
The counterpart of the
.I mmap
field, but enables including data mmap events
in the ring-buffer.
.TP
.IR "sample_id_all" " (Since Linux 2.6.38)"
If set, then TID, TIME, ID, CPU, and STREAM_ID can
additionally be included in
.RB non- PERF_RECORD_SAMPLE s
if the corresponding
.I sample_type
is selected.
.TP
.IR "exclude_host" " (Since Linux 3.2)"
Do not measure time spent in VM host
.TP
.IR "exclude_guest" " (Since Linux 3.2)"
Do not measure time spent in VM guest
.TP
.IR "exclude_callchain_kernel" " (Since Linux 3.7)"
Do not include kernel callchains.
.TP
.IR "exclude_callchain_user" " (Since Linux 3.7)"
Do not include user callchains.
.TP
.IR "wakeup_events" ", " "wakeup_watermark"
This union sets how many samples
.RI ( wakeup_events )
or bytes
.RI ( wakeup_watermark )
happen before an overflow signal happens.
Which one is used is selected by the
.I watermark
bitflag.

.I wakeup_events
only counts
.B PERF_RECORD_SAMPLE
record types.
To  receive a signal for every incoming
.B PERF_RECORD
type set
.I wakeup_watermark
to 1.
.TP
.IR "bp_type" " (Since Linux 2.6.33)"
This chooses the breakpoint type.
It is one of:
.RS
.TP
.BR HW_BREAKPOINT_EMPTY
no breakpoint
.TP
.BR HW_BREAKPOINT_R
count when we read the memory location
.TP
.BR HW_BREAKPOINT_W
count when we write the memory location
.TP
.BR HW_BREAKPOINT_RW
count when we read or write the memory location
.TP
.BR HW_BREAKPOINT_X
count when we execute code at the memory location
.LP
The values can be combined via a bitwise or, but the
combination of
.B HW_BREAKPOINT_R
or
.B HW_BREAKPOINT_W
with
.B HW_BREAKPOINT_X
is not allowed.
.RE
.TP
.IR "bp_addr" " (Since Linux 2.6.33)"
.I bp_addr
address of the breakpoint.
For execution breakpoints this is the memory address of the instruction
of interest; for read and write breakpoints it is the memory address
of the memory location of interest.
.TP
.IR "config1" " (Since Linux 2.6.39)"
.I config1
is used for setting events that need an extra register or otherwise
do not fit in the regular config field.
Raw OFFCORE_EVENTS on Nehalem/Westmere/SandyBridge use this field
on 3.3 and later kernels.
.TP
.IR "bp_len" " (Since Linux 2.6.33)"
.I bp_len
is the length of the breakpoint being measured if
.I type
is
.BR PERF_TYPE_BREAKPOINT .
Options are
.BR HW_BREAKPOINT_LEN_1 ,
.BR HW_BREAKPOINT_LEN_2 ,
.BR HW_BREAKPOINT_LEN_4 ,
.BR HW_BREAKPOINT_LEN_8 .
For an execution breakpoint, set this to
.IR sizeof(long) .
.TP
.IR "config2" " (Since Linux 2.6.39)"

.I config2
is a further extension of the
.I config1
field.
.TP
.IR "branch_sample_type" " (Since Linux 3.4)"
If
.B PERF_SAMPLE_BRANCH_STACK
is enabled, then this specifies what branches to include
in the branch record.
If the user does not set privilege level explicitly, the kernel
will use the event's privilege level.
Event and branch privilege levels do not have to match.
The value is formed by ORing together zero or more of the following values,
although
.B PERF_SAMPLE_BRANCH_ANY
covers all branch types.
.RS
.TP
.B PERF_SAMPLE_BRANCH_USER
Branch target is in user space
.TP
.B PERF_SAMPLE_BRANCH_KERNEL
Branch target is in kernel space
.TP
.B PERF_SAMPLE_BRANCH_HV
Branch target is in hypervisor
.TP
.B PERF_SAMPLE_BRANCH_ANY
Any branch type.
.TP
.B PERF_SAMPLE_BRANCH_ANY_CALL
Any call branch
.TP
.B PERF_SAMPLE_BRANCH_ANY_RETURN
Any return branch
.TP
.BR PERF_SAMPLE_BRANCH_IND_CALL
Indirect calls
.TP
.BR PERF_SAMPLE_BRANCH_PLM_ALL
User, kernel, and hv
.RE
.TP
.IR "sample_regs_user" " (Since Linux 3.7)"
This bitmask defines the set of user CPU registers to dump on samples.
The layout of the register mask is architecture specific and
described in the kernel header
.IR arch/ARCH/include/uapi/asm/perf_regs.h .
.TP
.IR "sample_stack_user" " (Since Linux 3.7)"
This defines the size of the user stack to dump if
.B PERF_SAMPLE_STACK_USER
is specified.
.SS Reading results
Once a
.BR perf_event_open ()
file descriptor  has been opened, the values
of the events can be read from the file descriptor.
The values that are there are specified by the
.I read_format
field in the
.I attr
structure at open time.

If you attempt to read into a buffer that is not big enough to hold the
data
.B ENOSPC
is returned

Here is the layout of the data returned by a read:
.IP * 2
If
.B PERF_FORMAT_GROUP
was specified to allow reading all events in a group at once:

.in +4n
.nf
struct read_format {
    u64 nr;            /* The number of events */
    u64 time_enabled;  /* if PERF_FORMAT_TOTAL_TIME_ENABLED */
    u64 time_running;  /* if PERF_FORMAT_TOTAL_TIME_RUNNING */
    struct
        u64 value;     /* The value of the event */
        u64 id;        /* if PERF_FORMAT_ID */
    } values[nr];
};
.fi
.in
.IP *
If
.B PERF_FORMAT_GROUP
was
.I not
specified:

.in +4n
.nf
struct read_format {
    u64 value;         /* The value of the event */
    u64 time_enabled;  /* if PERF_FORMAT_TOTAL_TIME_ENABLED */
    u64 time_running;  /* if PERF_FORMAT_TOTAL_TIME_RUNNING */
    u64 id;            /* if PERF_FORMAT_ID */
};
.fi
.in
.PP
The values read are as follows:
.TP
.I nr
The number of events in this file descriptor.
Only available if
.B PERF_FORMAT_GROUP
was specified.
.TP
.IR time_enabled ", " time_running
Total time the event was enabled and running.
Normally these are the same.
If more events are started
than available counter slots on the PMU, then multiplexing
happens and events run only part of the time.
In that case the
.I time_enabled
and
.I time running
values can be used to scale an estimated value for the count.
.TP
.I value
An unsigned 64-bit value containing the counter result.
.TP
.I id
A globally unique value for this particular event, only there if
.B PERF_FORMAT_ID
was specified in
.IR read_format .
.SS MMAP layout
When using
.BR perf_event_open ()
in sampled mode, asynchronous events
(like counter overflow or
.B PROT_EXEC
mmap tracking)
are logged into a ring-buffer.
This ring-buffer is created and accessed through
.BR mmap (2).

The mmap size should be 1+2^n pages, where the first page is a
metadata page
.RI ( "struct perf_event_mmap_page" )
that contains various
bits of information such as where the ring-buffer head is.

Before kernel 2.6.39, there is a bug that means you must allocate a mmap
ring buffer when sampling even if you do not plan to access it.

The structure of the first metadata mmap page is as follows:

.in +4n
.nf
struct perf_event_mmap_page {
    __u32 version;          /* version number of this structure */
    __u32 compat_version;   /* lowest version this is compat with */
    __u32 lock;             /* seqlock for synchronization */
    __u32 index;            /* hardware counter identifier */
    __s64 offset;           /* add to hardware counter value */
    __u64 time_enabled;     /* time event active */
    __u64 time_running;     /* time event on CPU */
    union {
        __u64   capabilities;
        __u64   cap_usr_time  : 1,
                cap_usr_rdpmc : 1,
    };
    __u16   pmc_width;
    __u16   time_shift;
    __u32   time_mult;
    __u64   time_offset;
    __u64   __reserved[120];   /* Pad to 1k */
    __u64   data_head;         /* head in the data section */
    __u64   data_tail;         /* user-space written tail */
}
.fi
.in

The following looks at the fields in the
.I perf_event_mmap_page
structure in more detail:
.TP
.I version
Version number of this structure.
.TP
.I compat_version
The lowest version this is compatible with.
.TP
.I lock
A seqlock for synchronization.
.TP
.I index
A unique hardware counter identifier.
.TP
.I offset
.\" FIXME clarify
Add this to hardware counter value??
.TP
.I time_enabled
Time the event was active.
.TP
.I time_running
Time the event was running.
.TP
.I cap_usr_time
User time capability
.TP
.I cap_usr_rdpmc
If the hardware supports user-space read of performance counters
without syscall (this is the "rdpmc" instruction on x86), then
the following code can be used to do a read:

.in +4n
.nf
u32 seq, time_mult, time_shift, idx, width;
u64 count, enabled, running;
u64 cyc, time_offset;
s64 pmc = 0;

do {
    seq = pc\->lock;
    barrier();
    enabled = pc\->time_enabled;
    running = pc\->time_running;

    if (pc\->cap_usr_time && enabled != running) {
        cyc = rdtsc();
        time_offset = pc\->time_offset;
        time_mult   = pc\->time_mult;
        time_shift  = pc\->time_shift;
    }

    idx = pc\->index;
    count = pc\->offset;

    if (pc\->cap_usr_rdpmc && idx) {
        width = pc\->pmc_width;
        pmc = rdpmc(idx \- 1);
    }

    barrier();
} while (pc\->lock != seq);
.fi
.in
.TP
.I pmc_width
If
.IR cap_usr_rdpmc ,
this field provides the bit-width of the value
read using the rdpmc or equivalent instruction.
This can be used to sign extend the result like:

.in +4n
.nf
pmc <<= 64 \- pmc_width;
pmc >>= 64 \- pmc_width; // signed shift right
count += pmc;
.fi
.in
.TP
.IR time_shift ", " time_mult ", " time_offset

If
.IR cap_usr_time ,
these fields can be used to compute the time
delta since time_enabled (in nanoseconds) using rdtsc or similar.
.nf

    u64 quot, rem;
    u64 delta;
    quot = (cyc >> time_shift);
    rem = cyc & ((1 << time_shift) \- 1);
    delta = time_offset + quot * time_mult +
            ((rem * time_mult) >> time_shift);
.fi

Where
.IR time_offset ,
.IR time_mult ,
.IR time_shift ,
and
.IR cyc
are read in the
seqcount loop described above.
This delta can then be added to
enabled and possible running (if idx), improving the scaling:
.nf

    enabled += delta;
    if (idx)
        running += delta;
    quot = count / running;
    rem  = count % running;
    count = quot * enabled + (rem * enabled) / running;
.fi
.TP
.I data_head
This points to the head of the data section.
The value continuously increases, it does not wrap.
The value needs to be manually wrapped by the size of the mmap buffer
before accessing the samples.

On SMP-capable platforms, after reading the data_head value,
user space should issue an rmb().
.TP
.I data_tail;
When the mapping is
.BR PROT_WRITE ,
the
.I data_tail
value should be written by user space to reflect the last read data.
In this case the kernel will not over-write unread data.
.PP
The following 2^n ring-buffer pages have the layout described below.

If
.I perf_event_attr.sample_id_all
is set, then all event types will
have the sample_type selected fields related to where/when (identity)
an event took place (TID, TIME, ID, CPU, STREAM_ID) described in
.B PERF_RECORD_SAMPLE
below, it will be stashed just after the
.I perf_event_header
and the fields already present for the existing
fields, i.e., at the end of the payload.
That way a newer perf.data
file will be supported by older perf tools, with these new optional
fields being ignored.

The mmap values start with a header:

.in +4n
.nf
struct perf_event_header {
    __u32   type;
    __u16   misc;
    __u16   size;
};
.fi
.in

Below, we describe the
.I perf_event_header
fields in more detail.
.TP
.I type
The
.I type
value is one of the below.
The values in the corresponding record (that follows the header)
depend on the
.I type
selected as shown.
.RS
.TP 4
.B PERF_RECORD_MMAP
The MMAP events record the
.B PROT_EXEC
mappings so that we can correlate
user-space IPs to code.
They have the following structure:

.in +4n
.nf
struct {
    struct perf_event_header header;
    u32    pid, tid;
    u64    addr;
    u64    len;
    u64    pgoff;
    char   filename[];
};
.fi
.in
.TP
.B PERF_RECORD_LOST
This record indicates when events are lost.

.in +4n
.nf
struct {
    struct perf_event_header header;
    u64 id;
    u64 lost;
};
.fi
.in
.RS
.TP
.I id
is the unique event ID for the samples that were lost.
.TP
.I lost
is the number of events that were lost.
.RE
.TP
.B PERF_RECORD_COMM
This record indicates a change in the process name.

.in +4n
.nf
struct {
    struct perf_event_header header;
    u32 pid, tid;
    char comm[];
};
.fi
.in
.TP
.B PERF_RECORD_EXIT
This record indicates a process exit event.

.in +4n
.nf
struct {
    struct perf_event_header header;
    u32 pid, ppid;
    u32 tid, ptid;
    u64 time;
};
.fi
.in
.TP
.BR PERF_RECORD_THROTTLE ", " PERF_RECORD_UNTHROTTLE
This record indicates a throttle/unthrottle event.

.in +4n
.nf
struct {
    struct perf_event_header header;
    u64 time;
    u64 id;
    u64 stream_id;
};
.fi
.in
.TP
.B PERF_RECORD_FORK
This record indicates a fork event.

.in +4n
.nf
struct {
    struct perf_event_header header;
    u32 pid, ppid;
    u32 tid, ptid;
    u64 time;
};
.fi
.in
.TP
.B PERF_RECORD_READ
This record indicates a read event.

.in +4n
.nf
struct {
    struct perf_event_header header;
    u32 pid, tid;
    struct read_format values;
};
.fi
.in
.TP
.B PERF_RECORD_SAMPLE
This record indicates a sample.

.in +4n
.nf
struct {
    struct perf_event_header header;
    u64   ip;         /* if PERF_SAMPLE_IP */
    u32   pid, tid;   /* if PERF_SAMPLE_TID */
    u64   time;       /* if PERF_SAMPLE_TIME */
    u64   addr;       /* if PERF_SAMPLE_ADDR */
    u64   id;         /* if PERF_SAMPLE_ID */
    u64   stream_id;  /* if PERF_SAMPLE_STREAM_ID */
    u32   cpu, res;   /* if PERF_SAMPLE_CPU */
    u64   period;     /* if PERF_SAMPLE_PERIOD */
    struct read_format v; /* if PERF_SAMPLE_READ */
    u64   nr;         /* if PERF_SAMPLE_CALLCHAIN */
    u64   ips[nr];    /* if PERF_SAMPLE_CALLCHAIN */
    u32   size;       /* if PERF_SAMPLE_RAW */
    char  data[size]; /* if PERF_SAMPLE_RAW */
    u64   bnr;        /* if PERF_SAMPLE_BRANCH_STACK */
    struct perf_branch_entry lbr[bnr];
                      /* if PERF_SAMPLE_BRANCH_STACK */
    u64   abi;        /* if PERF_SAMPLE_REGS_USER */
    u64   regs[weight(mask)];
                      /* if PERF_SAMPLE_REGS_USER */
    u64   size;       /* if PERF_SAMPLE_STACK_USER */
    char  data[size]; /* if PERF_SAMPLE_STACK_USER */
    u64   dyn_size;   /* if PERF_SAMPLE_STACK_USER */
    u64   weight;     /* if PERF_SAMPLE_WEIGHT */
    u64   data_src;   /* if PERF_SAMPLE_DATA_SRC */
};
.fi
.RS
.TP
.I ip
If
.B PERF_SAMPLE_IP
is enabled, then a 64-bit instruction
pointer value is included.
.TP
.IR pid ", " tid
If
.B PERF_SAMPLE_TID
is enabled, then a 32-bit process ID
and 32-bit thread ID are included.
.TP
.I time
If
.B PERF_SAMPLE_TIME
is enabled, then a 64-bit timestamp
is included.
This is obtained via local_clock() which is a hardware timestamp
if available and the jiffies value if not.
.TP
.I addr
If
.B PERF_SAMPLE_ADDR
is enabled, then a 64-bit address is included.
This is usually the address of a tracepoint,
breakpoint, or software event; otherwise the value is 0.
.TP
.I id
If
.B PERF_SAMPLE_ID
is enabled, a 64-bit unique ID is included.
If the event is a member of an event group, the group leader ID is returned.
This ID is the same as the one returned by
.BR PERF_FORMAT_ID .
.TP
.I stream_id
If
.B PERF_SAMPLE_STREAM_ID
is enabled, a 64-bit unique ID is included.
Unlike
.B PERF_SAMPLE_ID
the actual ID is returned, not the group leader.
This ID is the same as the one returned by
.BR PERF_FORMAT_ID .
.TP
.IR cpu ", " res
If
.B PERF_SAMPLE_CPU
is enabled, this is a 32-bit value indicating
which CPU was being used, in addition to a reserved (unused)
32-bit value.
.TP
.I period
If
.B PERF_SAMPLE_PERIOD
is enabled, a 64-bit value indicating
the current sampling period is written.
.TP
.I v
If
.B PERF_SAMPLE_READ
is enabled, a structure of type read_format
is included which has values for all events in the event group.
The values included depend on the
.I read_format
value used at
.BR perf_event_open ()
time.
.TP
.IR nr ", " ips[nr]
If
.B PERF_SAMPLE_CALLCHAIN
is enabled, then a 64-bit number is included
which indicates how many following 64-bit instruction pointers will
follow.
This is the current callchain.
.TP
.IR size ", " data[size]
If
.B PERF_SAMPLE_RAW
is enabled, then a 32-bit value indicating size
is included followed by an array of 8-bit values of length size.
The values are padded with 0 to have 64-bit alignment.

This RAW record data is opaque with respect to the ABI.
The ABI doesn't make any promises with respect to the stability
of its content, it may vary depending
on event, hardware, and kernel version.
.TP
.IR bnr ", " lbr[bnr]
If
.B PERF_SAMPLE_BRANCH_STACK
is enabled, then a 64-bit value indicating
the number of records is included, followed by
.I bnr
.I perf_branch_entry
structures which each include the fields:
.RS
.TP
.I from
indicating the source instruction (may not be a branch)
.TP
.I to
the branch target
.TP
.I mispred
the branch target was mispredicted
.TP
.I predicted
the branch target was predicted.
.RE
The entries are from most to least recent, so the first entry
has the most recent branch.

Support for
.I mispred
and
.I predicted
is optional; if not supported, both
values will be 0.

.TP
.IR abi ", " regs[weight(mask)]
If
.B PERF_SAMPLE_REGS_USER
is enabled, then the user CPU registers are recorded.

The
.I abi
field is one of
.BR PERF_SAMPLE_REGS_ABI_NONE ", " PERF_SAMPLE_REGS_ABI_32 " or "
.BR PERF_SAMPLE_REGS_ABI_64 .

The
.I regs
field is an array of the CPU registers that were specified by
the
.I sample_regs_user
attr field.
The number of values is the number of bits set in the
.I sample_regs_user
bitmask.
.TP
.IR size ", " data[size] ", " dyn_size
If
.B PERF_SAMPLE_STACK_USER
is enabled, then record the user stack to enable backtracing.
.I size
is the size requested by the user in
.I stack_user_size
or else the maximum record size.
.I data
is the stack data.
.I dyn_size
is the amount of data actually dumped (can be less than
.IR size ).
.TP
.I weight
If
.B PERF_SAMPLE_WEIGHT
is enabled, then a 64 bit value provided by the hardware
is recorded that indicates how costly the event was.
This allows expensive events to stand out more clearly
in profiles.
.TP
.I data_src
If
.B PERF_SAMPLE_DATA_SRC
is enabled, then a 64 bit value is recorded that is made up of
the following fields:
.RS
.TP
.I mem_op
type of opcode, a bitwise combination of
.B PERF_MEM_OP_NA
(not available),
.B PERF_MEM_OP_LOAD
(load instruction),
.B PERF_MEM_OP_STORE
(store instruction),
.B PERF_MEM_OP_PFETCH
(prefetch), and
.B PERF_MEM_OP_EXEC
(executable code).
.TP
.I mem_lvl
memory hierarchy level hit or miss, a bitwise combination of
.B PERF_MEM_LVL_NA
(not available),
.B PERF_MEM_LVL_HIT
(hit),
.B PERF_MEM_LVL_MISS
(miss),
.B PERF_MEM_LVL_L1
(level 1 cache),
.B PERF_MEM_LVL_LFB
(line fill buffer),
.B PERF_MEM_LVL_L2
(level 2 cache),
.B PERF_MEM_LVL_L3
(level 3 cache),
.B PERF_MEM_LVL_LOC_RAM
(local DRAM),
.B PERF_MEM_LVL_REM_RAM1
(remote DRAM 1 hop),
.B PERF_MEM_LVL_REM_RAM2
(remote DRAM 2 hops),
.B PERF_MEM_LVL_REM_CCE1
(remote cache 1 hop),
.B PERF_MEM_LVL_REM_CCE2
(remote cache 2 hops),
.B PERF_MEM_LVL_IO
(I/O memory), and
.B PERF_MEM_LVL_UNC
(uncached memory).
.TP
.I mem_snoop
snoop mode, a bitwise combination of
.B PERF_MEM_SNOOP_NA
(not available),
.B PERF_MEM_SNOOP_NONE
(no snoop),
.B PERF_MEM_SNOOP_HIT
(snoop hit),
.B PERF_MEM_SNOOP_MISS
(snoop miss), and
.B PERF_MEM_SNOOP_HITM
(snoop hit modified).
.TP
.I mem_lock
lock instruction, a bitwise combination of
.B PERF_MEM_LOCK_NA
(not available) and
.B PERF_MEM_LOCK_LOCKED
(locked transaction).
.TP
.I mem_dtlb
tlb access hit or miss, a bitwise combination of
.B PERF_MEM_TLB_NA
(not available),
.B PERF_MEM_TLB_HIT
(hit),
.B PERF_MEM_TLB_MISS
(miss),
.B PERF_MEM_TLB_L1
(level 1 TLB),
.B PERF_MEM_TLB_L2
(level 2 TLB),
.B PERF_MEM_TLB_WK
(hardware walker), and
.B PERF_MEM_TLB_OS
(OS fault handler).
.RE
.RE
.RE
.TP
.I misc
The
.I misc
field contains additional information about the sample.

The CPU mode can be determined from this value by masking with
.B PERF_RECORD_MISC_CPUMODE_MASK
and looking for one of the following (note these are not
bit masks, only one can be set at a time):
.RS
.TP
.B PERF_RECORD_MISC_CPUMODE_UNKNOWN
Unknown CPU mode.
.TP
.B PERF_RECORD_MISC_KERNEL
Sample happened in the kernel.
.TP
.B PERF_RECORD_MISC_USER
Sample happened in user code.
.TP
.B PERF_RECORD_MISC_HYPERVISOR
Sample happened in the hypervisor.
.TP
.B PERF_RECORD_MISC_GUEST_KERNEL
Sample happened in the guest kernel.
.TP
.B PERF_RECORD_MISC_GUEST_USER
Sample happened in guest user code.
.RE

.RS
In addition, one of the following bits can be set:
.TP
.B PERF_RECORD_MISC_MMAP_DATA
This is set when the mapping is not executable;
otherwise the mapping is executable.
.TP
.B PERF_RECORD_MISC_EXACT_IP
This indicates that the content of
.B PERF_SAMPLE_IP
points
to the actual instruction that triggered the event.
See also
.IR perf_event_attr.precise_ip .
.TP
.B PERF_RECORD_MISC_EXT_RESERVED
This indicates there is extended data available (currently not used).
.RE
.TP
.I size
This indicates the size of the record.
.RE
.SS Signal overflow
Events can be set to deliver a signal when a threshold is crossed.
The signal handler is set up using the
.BR poll (2),
.BR select (2),
.BR epoll (2)
and
.BR fcntl (2),
system calls.

To generate signals, sampling must be enabled
.RI ( sample_period
must have a non-zero value).

There are two ways to generate signals.

The first is to set a
.I wakeup_events
or
.I wakeup_watermark
value that will generate a signal if a certain number of samples
or bytes have been written to the mmap ring buffer.
In this case a signal of type
.B POLL_IN
is sent.

The other way is by use of the
.B PERF_EVENT_IOC_REFRESH
ioctl.
This ioctl adds to a counter that decrements each time the event overflows.
When non-zero, a
.B POLL_IN
signal is sent on overflow, but
once the value reaches 0, a signal is sent of type
.B POLL_HUP
and
the underlying event is disabled.

Note: on newer kernels (definitely noticed with 3.2)
.\" FIXME(Vince) : Find out when this was introduced
a signal is provided for every overflow, even if
.I wakeup_events
is not set.
.SS rdpmc instruction
Starting with Linux 3.4 on x86, you can use the
.I rdpmc
instruction to get low-latency reads without having to enter the kernel.
Note that using
.I rdpmc
is not necessarily faster than other methods for reading event values.

Support for this can be detected with the
.I cap_usr_rdpmc
field in the mmap page; documentation on how
to calculate event values can be found in that section.
.SS perf_event ioctl calls
.PP
Various ioctls act on
.BR perf_event_open ()
file descriptors
.TP
.B PERF_EVENT_IOC_ENABLE
Enables the individual event or event group specified by the
file descriptor argument.

If the
.B PERF_IOC_FLAG_GROUP
bit is set in the ioctl argument, then all events in a group are
enabled, even if the event specified is not the group leader
(but see BUGS).
.TP
.B PERF_EVENT_IOC_DISABLE
Disables the individual counter or event group specified by the
file descriptor argument.

Enabling or disabling the leader of a group enables or disables the
entire group; that is, while the group leader is disabled, none of the
counters in the group will count.
Enabling or disabling a member of a group other than the leader
affects only that counter; disabling a non-leader
stops that counter from counting but doesn't affect any other counter.

If the
.B PERF_IOC_FLAG_GROUP
bit is set in the ioctl argument, then all events in a group are
disabled, even if the event specified is not the group leader
(but see BUGS).
.TP
.B PERF_EVENT_IOC_REFRESH
Non-inherited overflow counters can use this
to enable a counter for a number of overflows specified by the argument,
after which it is disabled.
Subsequent calls of this ioctl add the argument value to the current
count.
A signal with
.B POLL_IN
set will happen on each overflow until the
count reaches 0; when that happens a signal with
POLL_HUP
set is sent and the event is disabled.
Using an argument of 0 is considered undefined behavior.
.TP
.B PERF_EVENT_IOC_RESET
Reset the event count specified by the
file descriptor argument to zero.
This resets only the counts; there is no way to reset the
multiplexing
.I time_enabled
or
.I time_running
values.

If the
.B PERF_IOC_FLAG_GROUP
bit is set in the ioctl argument, then all events in a group are
reset, even if the event specified is not the group leader
(but see BUGS).
.TP
.B PERF_EVENT_IOC_PERIOD
IOC_PERIOD is the command to update the period; it
does not update the current period but instead defers until next.

The argument is a pointer to a 64-bit value containing the
desired new period.
.TP
.B PERF_EVENT_IOC_SET_OUTPUT
This tells the kernel to report event notifications to the specified
file descriptor rather than the default one.
The file descriptors must all be on the same CPU.

The argument specifies the desired file descriptor, or \-1 if
output should be ignored.
.TP
.BR PERF_EVENT_IOC_SET_FILTER " (Since Linux 2.6.33)"
This adds an ftrace filter to this event.

The argument is a pointer to the desired ftrace filter.
.SS Using prctl
A process can enable or disable all the event groups that are
attached to it using the
.BR prctl (2)
.B PR_TASK_PERF_EVENTS_ENABLE
and
.B PR_TASK_PERF_EVENTS_DISABLE
operations.
This applies to all counters on the current process, whether created by
this process or by another, and does not affect any counters that this
process has created on other processes.
It enables or disables only
the group leaders, not any other members in the groups.
.SS perf_event related configuration files
Files in
.I /proc/sys/kernel/
.RS 4
.TP
.I /proc/sys/kernel/perf_event_paranoid

The
.I perf_event_paranoid
file can be set to restrict access to the performance counters.

2 - only allow user-space measurements

1 - (default) allow both kernel and user measurements

0 - allow access to CPU-specific data but not raw tracepoint samples

\-1 - no restrictions

The existence of the
.I perf_event_paranoid
file is the official method for determining if a kernel supports
.BR perf_event_open ().
.TP
.I /proc/sys/kernel/perf_event_max_sample_rate

This sets the maximum sample rate.
Setting this too high can allow
users to sample at a rate that impacts overall machine performance
and potentially lock up the machine.
The default value is
100000 (samples per second).
.TP
.I /proc/sys/kernel/perf_event_mlock_kb

Maximum number of pages an unprivileged user can mlock (2) .
The default is 516 (kB).

.RE
Files in
.I /sys/bus/event_source/devices/
.RS 4
Since Linux 2.6.34 the kernel supports having multiple PMUs
available for monitoring.
Information on how to program these PMUs can be found under
.IR /sys/bus/event_source/devices/ .
Each subdirectory corresponds to a different PMU.
.TP
.IR /sys/bus/event_source/devices/*/type " (Since Linux 2.6.38)"
This contains an integer that can be used in the
.I type
field of perf_event_attr to indicate you wish to use this PMU.
.TP
.IR /sys/bus/event_source/devices/*/rdpmc " (Since Linux 3.4)"
If this file is 1, then direct user-space access to the
performance counter registers is allowed via the rdpmc instruction.
This can be disabled by echoing 0 to the file.
.TP
.IR /sys/bus/event_source/devices/*/format/ " (Since Linux 3.4)"
This sub-directory contains information on the architecture-specific
sub-fields available for programming the various
.I config
fields in the perf_event_attr struct.

The content of each file is the name of the config field, followed
by a colon, followed by a series of integer bit ranges separated by
commas.
For example, the file
.I event
may contain the value
.I config1:1,6-10,44
which indicates that event is an attribute that occupies bits 1,6-10, and 44
of perf_event_attr::config1.
.TP
.IR /sys/bus/event_source/devices/*/events/ " (Since Linux 3.4)"
This sub-directory contains files with pre-defined events.
The contents are strings describing the event settings
expressed in terms of the fields found in the previously mentioned
.I ./format/
directory.
These are not necessarily complete lists of all events supported by
a PMU, but usually a subset of events deemed useful or interesting.

The content of each file is a list of attribute names
separated by commas.
Each entry has an optional value (either hex or decimal).
If no value is specified than it is assumed to be a single-bit
field with a value of 1.
An example entry may look like this:
.IR event=0x2,inv,ldlat=3 .
.TP
.I /sys/bus/event_source/devices/*/uevent
This file is the standard kernel device interface
for injecting hotplug events.
.TP
.IR /sys/bus/event_source/devices/*/cpumask " (Since Linux 3.7)"
The
.I cpumask
file contains a comma-separated list of integers that
indicate a representative CPU number for each socket (package)
on the motherboard.
This is needed when setting up uncore or northbridge events, as
those PMUs present socket-wide events.
.RE
.SH RETURN VALUE
.BR perf_event_open ()
returns the new file descriptor, or \-1 if an error occurred
(in which case,
.I errno
is set appropriately).
.SH ERRORS
.TP
.B EINVAL
Returned if the specified event is not available.
.TP
.B ENOSPC
Prior to Linux 3.3, if there was not enough room for the event,
.B ENOSPC
was returned.
Linus did not like this, and this was changed to
.BR EINVAL .
.B ENOSPC
is still returned if you try to read results into
too small of a buffer.
.SH VERSION
.BR perf_event_open ()
was introduced in Linux 2.6.31 but was called
.BR perf_counter_open () .
It was renamed in Linux 2.6.32.
.SH CONFORMING TO
This
.BR perf_event_open ()
system call Linux- specific
and should not be used in programs intended to be portable.
.SH NOTES
Glibc does not provide a wrapper for this system call; call it using
.BR syscall (2).
See the example below.

The official way of knowing if
.BR perf_event_open ()
support is enabled is checking
for the existence of the file
.IR /proc/sys/kernel/perf_event_paranoid .
.SH BUGS
The
.B F_SETOWN_EX
option to
.BR fcntl (2)
is needed to properly get overflow signals in threads.
This was introduced in Linux 2.6.32.

Prior to Linux 2.6.33 (at least for x86) the kernel did not check
if events could be scheduled together until read time.
The same happens on all known kernels if the NMI watchdog is enabled.
This means to see if a given set of events works you have to
.BR perf_event_open (),
start, then read before you know for sure you
can get valid measurements.

Prior to Linux 2.6.34 event constraints were not enforced by the kernel.
In that case, some events would silently return "0" if the kernel
scheduled them in an improper counter slot.

Prior to Linux 2.6.34 there was a bug when multiplexing where the
wrong results could be returned.

Kernels from Linux 2.6.35 to Linux 2.6.39 can quickly crash the kernel if
"inherit" is enabled and many threads are started.

Prior to Linux 2.6.35,
.B PERF_FORMAT_GROUP
did not work with attached processes.

In older Linux 2.6 versions,
refreshing an event group leader refreshed all siblings,
and refreshing with a parameter of 0 enabled infinite refresh.
This behavior is unsupported and should not be relied on.

There is a bug in the kernel code between
Linux 2.6.36 and Linux 3.0 that ignores the
"watermark" field and acts as if a wakeup_event
was chosen if the union has a
non-zero value in it.

From Linux 2.6.31 to Linux 3.4, the
.B PERF_IOC_FLAG_GROUP
ioctl argument was broken and would repeatedly operate
on the event specified rather than iterating across
all sibling events in a group.

Always double-check your results!
Various generalized events have had wrong values.
For example, retired branches measured
the wrong thing on AMD machines until Linux 2.6.35.
.SH EXAMPLE
The following is a short example that measures the total
instruction count of a call to
.BR printf (3).
.nf

#include <stdlib.h>
#include <stdio.h>
#include <unistd.h>
#include <string.h>
#include <sys/ioctl.h>
#include <linux/perf_event.h>
#include <asm/unistd.h>

long
perf_event_open(struct perf_event_attr *hw_event, pid_t pid,
                int cpu, int group_fd, unsigned long flags)
{
    int ret;

    ret = syscall(__NR_perf_event_open, hw_event, pid, cpu,
                   group_fd, flags);
    return ret;
}

int
main(int argc, char **argv)
{
    struct perf_event_attr pe;
    long long count;
    int fd;

    memset(&pe, 0, sizeof(struct perf_event_attr));
    pe.type = PERF_TYPE_HARDWARE;
    pe.size = sizeof(struct perf_event_attr);
    pe.config = PERF_COUNT_HW_INSTRUCTIONS;
    pe.disabled = 1;
    pe.exclude_kernel = 1;
    pe.exclude_hv = 1;

    fd = perf_event_open(&pe, 0, \-1, \-1, 0);
    if (fd == \-1) {
       fprintf(stderr, "Error opening leader %llx\\n", pe.config);
       exit(EXIT_FAILURE);
    }

    ioctl(fd, PERF_EVENT_IOC_RESET, 0);
    ioctl(fd, PERF_EVENT_IOC_ENABLE, 0);

    printf("Measuring instruction count for this printf\\n");

    ioctl(fd, PERF_EVENT_IOC_DISABLE, 0);
    read(fd, &count, sizeof(long long));

    printf("Used %lld instructions\\n", count);

    close(fd);
}
.fi
.SH SEE ALSO
.BR fcntl (2),
.BR mmap (2),
.BR open (2),
.BR prctl (2),
.BR read (2)