perf_event_open.2: Document new perf_event_paranoid default

[thirdparty/man-pages.git] / 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 2015-12-05 "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
.I pid
and
.I cpu
arguments allow specifying which process and CPU to monitor:
.TP
.BR "pid == 0" " and " "cpu == \-1"
This measures the calling process/thread on any CPU.
.TP
.BR "pid == 0" " and " "cpu >= 0"
This measures the calling process/thread only
when running on the specified CPU.
.TP
.BR "pid > 0" " and " "cpu == \-1"
This measures the specified process/thread on any CPU.
.TP
.BR "pid > 0" " and " "cpu >= 0"
This measures the specified process/thread only
when running on the specified CPU.
.TP
.BR "pid == \-1" " and " "cpu >= 0"
This measures all processes/threads on the specified CPU.
This requires
.B CAP_SYS_ADMIN
capability or a
.I /proc/sys/kernel/perf_event_paranoid
value of less than 1.
.TP
.BR "pid == \-1" " and " "cpu == \-1"
This setting is invalid and will return an error.
.P
When
.I pid
is greater than zero, permission to perform this system call
is governed by a ptrace access mode
.B PTRACE_MODE_READ_REALCREDS
check; see
.BR ptrace (2).

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 file descriptor 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\(emadded, divided (to get ratios), and so on\(emwith 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_CLOEXEC " (since Linux 3.14)"
.\" commit a21b0b354d4ac39be691f51c53562e2c24443d9e
This flag enables the close-on-exec flag for the created
event file descriptor,
so that the file descriptor is automatically closed on
.BR execve (2).
Setting the close-on-exec flags at creation time, rather than later with
.BR fcntl (2),
avoids potential race conditions where the calling thread invokes
.BR perf_event_open ()
and
.BR fcntl (2)
at the same time as another thread calls
.BR fork (2)
then
.BR execve (2).
.TP
.BR PERF_FLAG_FD_NO_GROUP
This flag tells the event to ignore the
.IR group_fd
parameter except for the purpose of setting up output redirection
using the
.B PERF_FLAG_FD_OUTPUT
flag.
.TP
.BR PERF_FLAG_FD_OUTPUT " (broken since Linux 2.6.35)"
.\" commit ac9721f3f54b27a16c7e1afb2481e7ee95a70318
This flag re-routes the event's sampled output to instead
be included in the mmap buffer of the event specified by
.IR group_fd .
.TP
.BR PERF_FLAG_PID_CGROUP " (since Linux 2.6.39)"
.\" commit e5d1367f17ba6a6fed5fd8b74e4d5720923e0c25
This flag activates per-container system-wide monitoring.
A container
is an abstraction that isolates a set of resources for finer-grained
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 */
          mmap2          :  1,  /* include mmap with inode data */
          comm_exec      :  1,  /* flag comm events that are due to exec */
          use_clockid    :  1,  /* use clockid for time fields */

          __reserved_1   : 38;

    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 */
    __s32 clockid;              /* clock to use for time fields */
    __u64 sample_regs_intr;     /* regs to dump on samples */
    __u32 aux_watermark;        /* aux bytes before wakeup */
    __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)"
.\" commit 24f1e32c60c45c89a997c73395b69c8af6f0a84e
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 2.6.38,
.\" commit 2e80a82a49c4c7eca4e35734380f28298ba5db19
.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 subdirectory 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.
.\" commit cb5d76999029ae7a517cb07dfa732c1b5a934fc2
.\" this was added much later when PERF_ATTR_SIZE_VER2 happened
.\" but the actual attr_size had increased in 2.6.33
.B PERF_ATTR_SIZE_VER2
is 80 corresponding to the addition of branch sampling in Linux 3.4.
.\" commit cb5d76999029ae7a517cb07dfa732c1b5a934fc2
.B PERF_ATTR_SIZE_VER3
is 96 corresponding to the addition
of
.I sample_regs_user
and
.I sample_stack_user
in Linux 3.7.
.\" commit 1659d129ed014b715b0b2120e6fd929bdd33ed03
.B PERF_ATTR_SIZE_VER4
is 104 corresponding to the addition of
.I sample_regs_intr
in Linux 3.19.
.\" commit 60e2364e60e86e81bc6377f49779779e6120977f
.B PERF_ATTR_SIZE_VER5
is 112 corresponding to the addition of
.I aux_watermark
in Linux 4.1.
.\" commit 1a5941312414c71dece6717da9a0fa1303127afa
.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.

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 2.6.35, this used
the wrong event on AMD processors.
.\" commit f287d332ce835f77a4f5077d2c0ef1e3f9ea42d2
.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)"
.\" commit 8f62242246351b5a4bc0c1f00c0c7003edea128a
Stalled cycles during issue.
.TP
.BR PERF_COUNT_HW_STALLED_CYCLES_BACKEND  " (since Linux 3.0)"
.\" commit 8f62242246351b5a4bc0c1f00c0c7003edea128a
Stalled cycles during retirement.
.TP
.BR PERF_COUNT_HW_REF_CPU_CYCLES  " (since Linux 3.3)"
.\" commit c37e17497e01fc0f5d2d6feb5723b210b3ab8890
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.
.\" commit e49a5bd38159dfb1928fd25b173bc9de4bbadb21
.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)"
.\" commit f7d7986060b2890fc26db6ab5203efbd33aa2497
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)"
.\" commit f7d7986060b2890fc26db6ab5203efbd33aa2497
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.
.TP
.BR PERF_COUNT_SW_DUMMY " (since Linux 3.12)"
.\" commit fa0097ee690693006ab1aea6c01ad3c851b65c77
This is a placeholder event that counts nothing.
Informational sample record types such as mmap or comm
must be associated with an active event.
This dummy event allows gathering such records without requiring
a counting event.
.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 3.1)"
.\" commit 89d6c0b5bdbb1927775584dcf532d98b3efe1477
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" event is one that generates an overflow notification
every N events, where N is given by
.IR sample_period .
A sampling event has
.IR sample_period " > 0."
When an overflow occurs, requested data is recorded
in the mmap buffer.
The
.I sample_type
field controls what data is recorded on each overflow.

.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
.BR 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)"
.\" commit bce38cd53e5ddba9cb6d708c4ef3d04a4016ec7e
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)"
.\" commit 4018994f3d8785275ef0e7391b75c3462c029e56
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)"
.\" commit c5ebcedb566ef17bda7b02686e0d658a7bb42ee7
Records the user level stack, allowing stack unwinding.
.TP
.BR PERF_SAMPLE_WEIGHT " (since Linux 3.10)"
.\" commit c3feedf2aaf9ac8bad6f19f5d21e4ee0b4b87e9c
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)"
.\" commit d6be9ad6c960f43800a6f118932bc8a5a4eadcd1
Records the data source: where in the memory hierarchy
the data associated with the sampled instruction came from.
This is available only if the underlying hardware
supports this feature.
.TP
.BR PERF_SAMPLE_IDENTIFIER " (since Linux 3.12)"
.\" commit ff3d527cebc1fa3707c617bfe9e74f53fcfb0955
Places the
.B SAMPLE_ID
value in a fixed position in the record,
either at the beginning (for sample events) or at the end
(if a non-sample event).

This was necessary because a sample stream may have
records from various different event sources with different
.I sample_type
settings.
Parsing the event stream properly was not possible because the
format of the record was needed to find
.BR SAMPLE_ID ,
but
the format could not be found without knowing what
event the sample belonged to (causing a circular
dependency).

The
.B PERF_SAMPLE_IDENTIFIER
setting makes the event stream always parsable
by putting
.B SAMPLE_ID
in a fixed location, even though
it means having duplicate
.B SAMPLE_ID
values in records.
.TP
.BR PERF_SAMPLE_TRANSACTION " (since Linux 3.13)"
.\" commit fdfbbd07e91f8fe387140776f3fd94605f0c89e5
Records reasons for transactional memory abort events
(for example, from Intel TSX transactional memory support).

The
.I precise_ip
setting must be greater than 0 and a transactional memory abort
event must be measured or no values will be recorded.
Also note that some perf_event measurements, such as sampled
cycle counting, may cause extraneous aborts (by causing an
interrupt during a transaction).
.TP
.BR PERF_SAMPLE_REGS_INTR " (since Linux 3.19)"
.\" commit 60e2364e60e86e81bc6377f49779779e6120977f
Records a subset of the current CPU register state
as specified by
.IR sample_regs_intr .
Unlike
.B PERF_SAMPLE_REGS_USER
the register values will return kernel register
state if the overflow happened while kernel
code is running.
If the CPU supports hardware sampling of
register state (i.e. PEBS on Intel x86) and
.I precise_ip
is set higher than zero then the register
values returned are those captured by
hardware at the time of the sampled
instruction's retirement.
.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 .

When creating an event group, typically the group leader is initialized
with
.I disabled
set to 1 and any child events are initialized with
.I disabled
set to 0.
Despite
.I disabled
being 0, the child events will not start until the group leader
is enabled.
.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.

Note that many unexpected situations may prevent events with the
.I exclusive
bit set from ever running.
This includes any users running a system-wide
measurement as well as any kernel use of the performance counters
(including the commonly enabled NMI Watchdog Timer interface).
.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 generation of
.B PERF_RECORD_MMAP
samples for every
.BR mmap (2)
call that has
.B PROT_EXEC
set.
This allows tools to notice new executable code being mapped into
a program (dynamic shared libraries for example)
so that addresses can be mapped back to the original code.
.TP
.IR "comm"
The
.I comm
bit enables tracking of process command name as modified by the
.BR exec (2)
and
.BR prctl (PR_SET_NAME)
system calls as well as writing to
.IR /proc/self/comm .
If the
.I comm_exec
flag is also successfully set (possible since Linux 3.16),
.\" commit 82b897782d10fcc4930c9d4a15b175348fdd2871
then the misc flag
.B PERF_RECORD_MISC_COMM_EXEC
can be used to differentiate the
.BR exec (2)
case from the others.
.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 an overflow notification happen when we cross the
.I wakeup_watermark
boundary.
Otherwise, overflow notifications happen after
.I wakeup_events
samples.
.TP
.IR "precise_ip" " (since Linux 2.6.35)"
.\" commit ab608344bcbde4f55ec4cd911b686b0ce3eae076
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)"
.\" commit 3af9e859281bda7eb7c20b51879cf43aa788ac2e
The counterpart of the
.I mmap
field.
This enables generation of
.B PERF_RECORD_MMAP
samples for
.BR mmap (2)
calls that do not have
.B PROT_EXEC
set (for example data and SysV shared memory).
.TP
.IR "sample_id_all" " (since Linux 2.6.38)"
.\" commit c980d1091810df13f21aabbce545fd98f545bbf7
If set, then TID, TIME, ID, STREAM_ID, and CPU can
additionally be included in
.RB non- PERF_RECORD_SAMPLE s
if the corresponding
.I sample_type
is selected.

If
.B PERF_SAMPLE_IDENTIFIER
is specified, then an additional ID value is included
as the last value to ease parsing the record stream.
This may lead to the
.I id
value appearing twice.

The layout is described by this pseudo-structure:
.in +4n
.nf
struct sample_id {
    { u32 pid, tid; } /* if PERF_SAMPLE_TID set        */
    { u64 time;     } /* if PERF_SAMPLE_TIME set       */
    { u64 id;       } /* if PERF_SAMPLE_ID set         */
    { u64 stream_id;} /* if PERF_SAMPLE_STREAM_ID set  */
    { u32 cpu, res; } /* if PERF_SAMPLE_CPU set        */
    { u64 id;       } /* if PERF_SAMPLE_IDENTIFIER set */
};
.fi
.TP
.IR "exclude_host" " (since Linux 3.2)"
.\" commit a240f76165e6255384d4bdb8139895fac7988799
When conducting measurements that include processes running
VM instances (i.e. have executed a
.I KVM_RUN
.BR ioctl (2)
) only measure events happening inside a guest instance.
This is only meaningful outside the guests; this setting does
not change counts gathered inside of a guest.
Currently, this functionality is x86 only.
.TP
.IR "exclude_guest" " (since Linux 3.2)"
.\" commit a240f76165e6255384d4bdb8139895fac7988799
When conducting measurements that include processes running
VM instances (i.e. have executed a
.I KVM_RUN
.BR ioctl (2)
) do not measure events happening inside guest instances.
This is only meaningful outside the guests; this setting does
not change counts gathered inside of a guest.
Currently, this functionality is x86 only.
.TP
.IR "exclude_callchain_kernel" " (since Linux 3.7)"
.\" commit d077526485d5c9b12fe85d0b2b3b7041e6bc5f91
Do not include kernel callchains.
.TP
.IR "exclude_callchain_user" " (since Linux 3.7)"
.\" commit d077526485d5c9b12fe85d0b2b3b7041e6bc5f91
Do not include user callchains.
.TP
.IR "mmap2" " (since Linux 3.16)"
.\" commit 13d7a2410fa637f450a29ecb515ac318ee40c741
.\" This is tricky; was committed during 3.12 development
.\" but right before release was disabled.
.\" So while you could select mmap2 starting with 3.12
.\" it did not work until 3.16
.\" commit a5a5ba72843dd05f991184d6cb9a4471acce1005
Generate an extended executable mmap record that contains enough
additional information to uniquely identify shared mappings.
The
.I mmap
flag must also be set for this to work.
.TP
.IR "comm_exec" " (since Linux 3.16)"
.\" commit 82b897782d10fcc4930c9d4a15b175348fdd2871
This is purely a feature-detection flag, it does not change
kernel behavior.
If this flag can successfully be set, then, when
.I comm
is enabled, the
.B PERF_RECORD_MISC_COMM_EXEC
flag will be set in the
.I misc
field of a comm record header if the rename event being
reported was caused by a call to
.BR exec (2).
This allows tools to distinguish between the various
types of process renaming.
.TP
.IR "use_clockid" " (since Linux 4.1)"
.\" commit 34f439278cef7b1177f8ce24f9fc81dfc6221d3b
This allows selecting which internal Linux clock to use
when generating timestamps via the
.I clockid
field.
This can make it easier to correlate perf sample times with
timestamps generated by other tools.
.TP
.IR "wakeup_events" ", " "wakeup_watermark"
This union sets how many samples
.RI ( wakeup_events )
or bytes
.RI ( wakeup_watermark )
happen before an overflow notification happens.
Which one is used is selected by the
.I watermark
bit flag.

.I wakeup_events
counts only
.B PERF_RECORD_SAMPLE
record types.
To receive overflow notification for all
.B PERF_RECORD
types choose watermark and set
.I wakeup_watermark
to 1.

Prior to Linux 3.0 setting
.\" commit f506b3dc0ec454a16d40cab9ee5d75435b39dc50
.I wakeup_events
to 0 resulted in no overflow notifications;
more recent kernels treat 0 the same as 1.
.TP
.IR "bp_type" " (since Linux 2.6.33)"
.\" commit 24f1e32c60c45c89a997c73395b69c8af6f0a84e
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)"
.\" commit 24f1e32c60c45c89a997c73395b69c8af6f0a84e
.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)"
.\" commit a7e3ed1e470116c9d12c2f778431a481a6be8ab6
.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)"
.\" commit 24f1e32c60c45c89a997c73395b69c8af6f0a84e
.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)"
.\" commit a7e3ed1e470116c9d12c2f778431a481a6be8ab6

.I config2
is a further extension of the
.I config1
field.
.TP
.IR "branch_sample_type" " (since Linux 3.4)"
.\" commit bce38cd53e5ddba9cb6d708c4ef3d04a4016ec7e
If
.B PERF_SAMPLE_BRANCH_STACK
is enabled, then this specifies what branches to include
in the branch record.

The first part of the value is the privilege level, which
is a combination of one of the following values.
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.
.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_PLM_ALL
A convenience value that is the three preceding values ORed together.
.P
In addition to the privilege value, at least one or more of the
following bits must be set.
.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
.B PERF_SAMPLE_BRANCH_IND_CALL
Indirect calls.
.TP
.BR PERF_SAMPLE_BRANCH_COND " (since Linux 3.16)"
.\" commit bac52139f0b7ab31330e98fd87fc5a2664951050
Conditional branches.
.TP
.BR PERF_SAMPLE_BRANCH_ABORT_TX " (since Linux 3.11)"
.\" commit 135c5612c460f89657c4698fe2ea753f6f667963
Transactional memory aborts.
.TP
.BR PERF_SAMPLE_BRANCH_IN_TX " (since Linux 3.11)"
.\" commit 135c5612c460f89657c4698fe2ea753f6f667963
Branch in transactional memory transaction.
.TP
.BR PERF_SAMPLE_BRANCH_NO_TX " (since Linux 3.11)"
.\" commit 135c5612c460f89657c4698fe2ea753f6f667963
Branch not in transactional memory transaction.
.BR PERF_SAMPLE_BRANCH_CALL_STACK " (since Linux 4.1)"
.\" commit 2c44b1936bb3b135a3fac8b3493394d42e51cf70
Branch is part of a hardware-generated call stack.
This requires hardware support, currently only found
on Intel x86 Haswell or newer.
.RE
.TP
.IR "sample_regs_user" " (since Linux 3.7)"
.\" commit 4018994f3d8785275ef0e7391b75c3462c029e56
This bit mask 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)"
.\" commit c5ebcedb566ef17bda7b02686e0d658a7bb42ee7
This defines the size of the user stack to dump if
.B PERF_SAMPLE_STACK_USER
is specified.
.TP
.IR "clockid" " (since Linux 4.1)"
.\" commit 34f439278cef7b1177f8ce24f9fc81dfc6221d3b
If
.I use_clockid
is set, then this field selects which internal Linux timer to
use for timestamps.
The available timers are defined in
.IR linux/time.h ,
with
.BR CLOCK_MONOTONIC ,
.BR CLOCK_MONOTONIC_RAW ,
.BR CLOCK_REALTIME ,
.BR CLOCK_BOOTTIME ,
and
.B CLOCK_TAI
currently supported.
.TP
.IR "aux_watermark" " (since Linux 4.1)"
.\" commit 1a5941312414c71dece6717da9a0fa1303127afa
This specifies how much data is required to trigger a
.B PERF_RECORD_AUX
sample.
.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,
then 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 present 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 an 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;
        struct {
            __u64 cap_usr_time / cap_usr_rdpmc / cap_bit0 : 1,
                  cap_bit0_is_deprecated : 1,
                  cap_user_rdpmc         : 1,
                  cap_user_time          : 1,
                  cap_user_time_zero     : 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 */
    __u64 data_offset;       /* where the buffer starts */
    __u64 data_size;         /* data buffer size */
    __u64 aux_head;
    __u64 aux_tail;
    __u64 aux_offset;
    __u64 aux_size;

}
.fi
.in

The following list describes 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
When using rdpmc for reads this offset value
must be added to the one returned by rdpmc to get
the current total event count.
.TP
.I time_enabled
Time the event was active.
.TP
.I time_running
Time the event was running.
.TP
.IR cap_usr_time " / " cap_usr_rdpmc " / " cap_bit0 " (since Linux 3.4)"
.\" commit c7206205d00ab375839bd6c7ddb247d600693c09
There was a bug in the definition of
.I cap_usr_time
and
.I cap_usr_rdpmc
from Linux 3.4 until Linux 3.11.
Both bits were defined to point to the same location, so it was
impossible to know if
.I cap_usr_time
or
.I cap_usr_rdpmc
were actually set.

Starting with Linux 3.12, these are renamed to
.\" commit fa7315871046b9a4c48627905691dbde57e51033
.I cap_bit0
and you should use the
.I cap_user_time
and
.I cap_user_rdpmc
fields instead.
.TP
.IR cap_bit0_is_deprecated " (since Linux 3.12)"
.\" commit fa7315871046b9a4c48627905691dbde57e51033
If set, this bit indicates that the kernel supports
the properly separated
.I cap_user_time
and
.I cap_user_rdpmc
bits.

If not-set, it indicates an older kernel where
.I cap_usr_time
and
.I cap_usr_rdpmc
map to the same bit and thus both features should
be used with caution.
.TP
.IR cap_user_rdpmc " (since Linux 3.12)"
.\" commit fa7315871046b9a4c48627905691dbde57e51033
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;

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;
        count += rdpmc(idx \- 1);
    }

    barrier();
} while (pc\->lock != seq);
.fi
.in
.TP
.IR cap_user_time " (since Linux 3.12)"
.\" commit fa7315871046b9a4c48627905691dbde57e51033
This bit indicates the hardware has a constant, nonstop
timestamp counter (TSC on x86).
.TP
.IR cap_user_time_zero " (since Linux 3.12)"
.\" commit fa7315871046b9a4c48627905691dbde57e51033
Indicates the presence of
.I time_zero
which allows mapping timestamp values to
the hardware clock.
.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
.IR time_zero " (since Linux 3.12)"
.\" commit fa7315871046b9a4c48627905691dbde57e51033

If
.I cap_usr_time_zero
is set, then the hardware clock (the TSC timestamp counter on x86)
can be calculated from the
.IR time_zero ", " time_mult ", and " time_shift " values:"

.nf
    time = timestamp - time_zero;
    quot = time / time_mult;
    rem  = time % time_mult;
    cyc = (quot << time_shift) + (rem << time_shift) / time_mult;
.fi

And vice versa:

.nf
    quot = cyc >> time_shift;
    rem  = cyc & ((1 << time_shift) - 1);
    timestamp = time_zero + quot * time_mult +
        ((rem * time_mult) >> time_shift);
.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
.I 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 overwrite unread data.
.TP
.IR data_offset " (since Linux 4.1)"
.\" commit e8c6deac69629c0cb97c3d3272f8631ef17f8f0f
Contains the offset of the location in the mmap buffer
where perf sample data begins.
.TP
.IR data_size " (since Linux 4.1)"
.\" commit e8c6deac69629c0cb97c3d3272f8631ef17f8f0f
Contains the size of the perf sample region within
the mmap buffer.
.TP
.IR aux_head ", " aux_tail ", " aux_offset ", " aux_size " (since Linux 4.1)
.\" commit 45bfb2e50471abbbfd83d40d28c986078b0d24ff
The AUX region allows mmaping a separate sample buffer for
high-bandwidth data streams (separate from the main perf sample buffer).
An example of a high-bandwidth stream is instruction tracing support,
as is found in newer Intel processors.

To set up an AUX area, first
.I aux_offset
needs to be set with an offset greater than
.IR data_offset + data_size
and
.I aux_size
needs to be set to the desired buffer size.
The desired offset and size must be page aligned, and the size
must be a power of two.
These values are then passed to mmap in order to map the AUX buffer.
Pages in the AUX buffer are included as part of the
.BR RLIMIT_MEMLOCK
resource limit (see
.BR setrlimit (2)),
and also as part of the
.I perf_event_mlock_kb
allowance.

By default, the AUX buffer will be truncated if it will not fit
in the available space in the ring buffer.
If the AUX buffer is mapped as a read only buffer, then it will
operate in ring buffer mode where old data will be overwritten
by new.
In overwrite mode, it might not be possible to infer where the
new data began, and it is the consumer's job to disable
measurement while reading to avoid possible data races.

The
.IR aux_head " and " aux_tail
ring buffer pointers have the same behavior and ordering
rules as the previous described
.IR data_head " and " data_tail .
.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, that is, 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.
For ease of reading,
the fields with shorter descriptions are presented first.
.TP
.I size
This indicates the size of the record.
.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
.BR PERF_RECORD_MISC_GUEST_KERNEL " (since Linux 2.6.35)"
.\" commit 39447b386c846bbf1c56f6403c5282837486200f
Sample happened in the guest kernel.
.TP
.B PERF_RECORD_MISC_GUEST_USER " (since Linux 2.6.35)"
.\" commit 39447b386c846bbf1c56f6403c5282837486200f
Sample happened in guest user code.
.RE

.RS
In addition, one of the following bits can be set:
.TP
.BR PERF_RECORD_MISC_MMAP_DATA " (since Linux 3.10)"
.\" commit 2fe85427e3bf65d791700d065132772fc26e4d75
This is set when the mapping is not executable;
otherwise the mapping is executable.
.TP
.BR PERF_RECORD_MISC_COMM_EXEC " (since Linux 3.16)"
.\" commit 82b897782d10fcc4930c9d4a15b175348fdd2871
This is set for a
.B PERF_RECORD_COMM
record on kernels more recent than Linux 3.16
if a process name change was caused by an
.BR exec (2)
system call.
It is an alias for
.B PERF_RECORD_MISC_MMAP_DATA
since the two values would not be set in the same record.
.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
.BR PERF_RECORD_MISC_EXT_RESERVED " (since Linux 2.6.35)"
.\" commit 1676b8a077c352085d52578fb4f29350b58b6e74
This indicates there is extended data available (currently not used).
.RE
.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
.RS
.TP
.I pid
is the process ID.
.TP
.I tid
is the thread ID.
.TP
.I addr
is the address of the allocated memory.
.I len
is the length of the allocated memory.
.I pgoff
is the page offset of the allocated memory.
.I filename
is a string describing the backing of the allocated memory.
.RE
.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;
    struct sample_id sample_id;
};
.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;
    u32 tid;
    char comm[];
    struct sample_id sample_id;
};
.fi
.in
.RS
.TP
.I pid
is the process ID.
.TP
.I tid
is the thread ID.
.TP
.I comm
is a string containing the new name of the process.
.RE
.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;
    struct sample_id sample_id;
};
.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;
    struct sample_id sample_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;
    struct sample_id sample_id;
};
.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;
    struct sample_id sample_id;
};
.fi
.in
.TP
.B PERF_RECORD_SAMPLE
This record indicates a sample.

.in +4n
.nf
struct {
    struct perf_event_header header;
    u64   sample_id;  /* if PERF_SAMPLE_IDENTIFIER */
    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 && size != 0 */
    u64   weight;     /* if PERF_SAMPLE_WEIGHT */
    u64   data_src;   /* if PERF_SAMPLE_DATA_SRC */
    u64   transaction;/* if PERF_SAMPLE_TRANSACTION */
    u64   abi;        /* if PERF_SAMPLE_REGS_INTR */
    u64   regs[weight(mask)];
                      /* if PERF_SAMPLE_REGS_INTR */
};
.fi
.RS 4
.TP 4
.I sample_id
If
.B PERF_SAMPLE_IDENTIFIER
is enabled, a 64-bit unique ID is included.
This is a duplication of the
.B PERF_SAMPLE_ID
.I id
value, but included at the beginning of the sample
so parsers can easily obtain the value.
.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
This indicates 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.
.TP
.IR in_tx " (since Linux 3.11)"
.\" commit 135c5612c460f89657c4698fe2ea753f6f667963
The branch was in a transactional memory transaction.
.TP
.IR abort " (since Linux 3.11)"
.\" commit 135c5612c460f89657c4698fe2ea753f6f667963
The branch was in an aborted transactional memory transaction.
.P
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.

The type of branches recorded is specified by the
.I branch_sample_type
field.
.RE
.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
bit mask.
.TP
.IR size ", " data[size] ", " dyn_size
If
.B PERF_SAMPLE_STACK_USER
is enabled, then the user stack is recorded.
This can be used to generate stack backtraces.
.I size
is the size requested by the user in
.I sample_stack_user
or else the maximum record size.
.I data
is the stack data (a raw dump of the memory pointed to by the
stack pointer at the time of sampling).
.I dyn_size
is the amount of data actually dumped (can be less than
.IR size ).
Note that
.I dyn_size
is omitted if
.I size
is 0.
.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 4
.I mem_op
Type of opcode, a bitwise combination of:

.PD 0
.RS
.TP 24
.B PERF_MEM_OP_NA
Not available
.TP
.B PERF_MEM_OP_LOAD
Load instruction
.TP
.B PERF_MEM_OP_STORE
Store instruction
.TP
.B PERF_MEM_OP_PFETCH
Prefetch
.TP
.B PERF_MEM_OP_EXEC
Executable code
.RE
.PD
.TP
.I mem_lvl
Memory hierarchy level hit or miss, a bitwise combination of
the following, shifted left by
.BR PERF_MEM_LVL_SHIFT :

.PD 0
.RS
.TP 24
.B PERF_MEM_LVL_NA
Not available
.TP
.B PERF_MEM_LVL_HIT
Hit
.TP
.B PERF_MEM_LVL_MISS
Miss
.TP
.B PERF_MEM_LVL_L1
Level 1 cache
.TP
.B PERF_MEM_LVL_LFB
Line fill buffer
.TP
.B PERF_MEM_LVL_L2
Level 2 cache
.TP
.B PERF_MEM_LVL_L3
Level 3 cache
.TP
.B PERF_MEM_LVL_LOC_RAM
Local DRAM
.TP
.B PERF_MEM_LVL_REM_RAM1
Remote DRAM 1 hop
.TP
.B PERF_MEM_LVL_REM_RAM2
Remote DRAM 2 hops
.TP
.B PERF_MEM_LVL_REM_CCE1
Remote cache 1 hop
.TP
.B PERF_MEM_LVL_REM_CCE2
Remote cache 2 hops
.TP
.B PERF_MEM_LVL_IO
I/O memory
.TP
.B PERF_MEM_LVL_UNC
Uncached memory
.RE
.PD
.TP
.I mem_snoop
Snoop mode, a bitwise combination of the following, shifted left by
.BR PERF_MEM_SNOOP_SHIFT :

.PD 0
.RS
.TP 24
.B PERF_MEM_SNOOP_NA
Not available
.TP
.B PERF_MEM_SNOOP_NONE
No snoop
.TP
.B PERF_MEM_SNOOP_HIT
Snoop hit
.TP
.B PERF_MEM_SNOOP_MISS
Snoop miss
.TP
.B PERF_MEM_SNOOP_HITM
Snoop hit modified
.RE
.PD
.TP
.I mem_lock
Lock instruction, a bitwise combination of the following, shifted left by
.BR PERF_MEM_LOCK_SHIFT :

.PD 0
.RS
.TP 24
.B PERF_MEM_LOCK_NA
Not available
.TP
.B PERF_MEM_LOCK_LOCKED
Locked transaction
.RE
.PD
.TP
.I mem_dtlb
TLB access hit or miss, a bitwise combination of the following, shifted
left by
.BR PERF_MEM_TLB_SHIFT :

.PD 0
.RS
.TP 24
.B PERF_MEM_TLB_NA
Not available
.TP
.B PERF_MEM_TLB_HIT
Hit
.TP
.B PERF_MEM_TLB_MISS
Miss
.TP
.B PERF_MEM_TLB_L1
Level 1 TLB
.TP
.B PERF_MEM_TLB_L2
Level 2 TLB
.TP
.B PERF_MEM_TLB_WK
Hardware walker
.TP
.B PERF_MEM_TLB_OS
OS fault handler
.RE
.PD
.RE
.TP
.I transaction
If the
.B PERF_SAMPLE_TRANSACTION
flag is set, then a 64-bit field is recorded describing
the sources of any transactional memory aborts.

The field is a bitwise combination of the following values:
.RS
.TP
.B PERF_TXN_ELISION
Abort from an elision type transaction (Intel-CPU-specific).
.TP
.B PERF_TXN_TRANSACTION
Abort from a generic transaction.
.TP
.B PERF_TXN_SYNC
Synchronous abort (related to the reported instruction).
.TP
.B PERF_TXN_ASYNC
Asynchronous abort (not related to the reported instruction).
.TP
.B PERF_TXN_RETRY
Retryable abort (retrying the transaction may have succeeded).
.TP
.B PERF_TXN_CONFLICT
Abort due to memory conflicts with other threads.
.TP
.B PERF_TXN_CAPACITY_WRITE
Abort due to write capacity overflow.
.TP
.B PERF_TXN_CAPACITY_READ
Abort due to read capacity overflow.
.RE
.IP
In addition, a user-specified abort code can be obtained from
the high 32 bits of the field by shifting right by
.B PERF_TXN_ABORT_SHIFT
and masking with
.BR PERF_TXN_ABORT_MASK .
.TP
.IR abi ", " regs[weight(mask)]
If
.B PERF_SAMPLE_REGS_INTR
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_intr
attr field.
The number of values is the number of bits set in the
.I sample_regs_intr
bit mask.
.RE
.TP
.B PERF_RECORD_MMAP2
This record includes extended information on
.BR mmap (2)
calls returning executable mappings.
The format is similar to that of the
.B PERF_RECORD_MMAP
record, but includes extra values that allow uniquely identifying
shared mappings.

.in +4n
.nf
struct {
    struct perf_event_header header;
    u32 pid;
    u32 tid;
    u64 addr;
    u64 len;
    u64 pgoff;
    u32 maj;
    u32 min;
    u64 ino;
    u64 ino_generation;
    u32 prot;
    u32 flags;
    char filename[];
    struct sample_id sample_id;
};
.fi
.RS
.TP
.I pid
is the process ID.
.TP
.I tid
is the thread ID.
.TP
.I addr
is the address of the allocated memory.
.TP
.I len
is the length of the allocated memory.
.TP
.I pgoff
is the page offset of the allocated memory.
.TP
.I maj
is the major ID of the underlying device.
.TP
.I min
is the minor ID of the underlying device.
.TP
.I ino
is the inode number.
.TP
.I ino_generation
is the inode generation.
.TP
.I prot
is the protection information.
.TP
.I flags
is the flags information.
.TP
.I filename
is a string describing the backing of the allocated memory.
.RE
.TP
.BR PERF_RECORD_AUX " (since Linux 4.1)"
\" commit 68db7e98c3a6ebe7284b6cf14906ed7c55f3f7f0
This record reports that new data is available in the separate
AUX buffer region.

.in +4n
.nf
struct {
    struct perf_event_header header;
    u64 aux_offset;
    u64 aux_size;
    u64 flags;
    struct sample_id sample_id;
};
.fi
.RS
.TP
.I aux_offset
offset in the AUX mmap region where the new data begins.
.TP
.I aux_size
size of the data made available.
.TP
.I flags
describes the AUX update.
.RS
.TP
.B PERF_AUX_FLAG_TRUNCATED
if set, then the data returned was truncated to fit the available
buffer size.
.TP
.B PERF_AUX_FLAG_OVERWRITE
.\" commit 2023a0d2829e521fe6ad6b9907f3f90bfbf57142
if set, then the data returned has overwritten previous data.
.RE
.RE
.TP
.BR PERF_RECORD_ITRACE_START " (since Linux 4.1)"
\" ec0d7729bbaed4b9d2d3fada693278e13a3d1368
This record indicates which process has initiated an instruction
trace event, allowing tools to properly correlate the instruction
addresses in the AUX buffer with the proper executable.

.in +4n
.nf
struct {
    struct perf_event_header header;
    u32 pid;
    u32 tid;
};
.fi
.RS
.TP
.I pid
process ID of the thread starting an instruction trace.
.TP
.I tid
thread ID of the thread starting an instruction trace.
.RE
.RE
.SS Overflow handling
Events can be set to notify when a threshold is crossed,
indicating an overflow.
Overflow conditions can be captured by monitoring the
event file descriptor with
.BR poll (2),
.BR select (2),
or
.BR epoll (2).
Alternately, a SIGIO signal handler can be created and
the event configured with
.BR fcntl (2)
to generate SIGIO signals.

Overflows are generated only by sampling events
.RI ( sample_period
must have a nonzero value).

There are two ways to generate overflow notifications.

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

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 nonzero,
.B POLL_IN
is indicated, but
once the counter reaches 0
.B POLL_HUP
is indicated and
the underlying event is disabled.

Refreshing an event group leader refreshes all siblings and
refreshing with a parameter of 0 currently enables infinite
refreshes;
these behaviors are unsupported and should not be relied on.
.\" See https://lkml.org/lkml/2011/5/24/337

Starting with Linux 3.18,
.\" commit 179033b3e064d2cd3f5f9945e76b0a0f0fbf4883
.B POLL_HUP
is indicated if the event being monitored is attached to a different
process and that process exits.
.SS rdpmc instruction
Starting with Linux 3.4 on x86, you can use the
.\" commit c7206205d00ab375839bd6c7ddb247d600693c09
.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.

Originally, when rdpmc support was enabled, any process (not just ones
with an active perf event) could use the rdpmc instruction to access
the counters.
Starting with Linux 4.0
.\" 7911d3f7af14a614617e38245fedf98a724e46a9
rdpmc support is only allowed if an event is currently enabled
in a process's context.
To restore the old behavior, write the value 2 to
.IR /sys/devices/cpu/rdpmc .
.SS perf_event ioctl calls
.PP
Various ioctls act on
.BR perf_event_open ()
file descriptors:
.TP
.B PERF_EVENT_IOC_ENABLE
This 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
This 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.
An overflow notification with
.B POLL_IN
set will happen on each overflow until the
count reaches 0; when that happens a notification with
.B 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
This updates the overflow period for the event.

Since Linux 3.7 (on ARM)
.\" commit 3581fe0ef37ce12ac7a4f74831168352ae848edc
and Linux 3.14 (all other architectures),
.\" commit bad7192b842c83e580747ca57104dd51fe08c223
the new period takes effect immediately.
On older kernels, the new period did not take effect until
after the next overflow.

The argument is a pointer to a 64-bit value containing the
desired new period.

Prior to Linux 2.6.36
.\" commit ad0cf3478de8677f720ee06393b3147819568d6a
this ioctl always failed due to a bug
in the kernel.
.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)"
.\" commit 6fb2915df7f0747d9044da9dbff5b46dc2e20830
This adds an ftrace filter to this event.

The argument is a pointer to the desired ftrace filter.
.TP
.BR PERF_EVENT_IOC_ID " (since Linux 3.12)"
.\" commit cf4957f17f2a89984915ea808876d9c82225b862
This returns the event ID value for the given event file descriptor.

The argument is a pointer to a 64-bit unsigned integer
to hold the result.
.TP
.BR PERF_EVENT_IOC_SET_BPF " (since Linux 4.1)"
.\" commit 2541517c32be2531e0da59dfd7efc1ce844644f5
This allows attaching a Berkeley Packet Filter (BPF)
program to an existing kprobe tracepoint event.
You need
.B CAP_SYS_ADMIN
privileges to use this ioctl.

The argument is a BPF program file descriptor that was created by
a previous
.BR bpf (2)
system call.
.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 calling 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

.\" default changed in commit 0161028b7c8aebef64194d3d73e43bc3b53b5c66
The
.I perf_event_paranoid
file can be set to restrict access to the performance counters.
.RS
.IP 2 4
allow only user-space measurements (default since Linux 4.6).
.IP 1
allow both kernel and user measurements (default before Linux 4.6).
.IP 0
allow access to CPU-specific data but not raw tracepoint samples.
.IP \-1
no restrictions.
.RE
.IP
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
.BR 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)"
.\" commit abe43400579d5de0078c2d3a760e6598e183f871
This contains an integer that can be used in the
.I type
field of
.I perf_event_attr
to indicate that you wish to use this PMU.
.TP
.IR /sys/bus/event_source/devices/cpu/rdpmc " (since Linux 3.4)"
.\" commit 0c9d42ed4cee2aa1dfc3a260b741baae8615744f
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.

As of Linux 4.0
.\" a66734297f78707ce39d756b656bfae861d53f62
.\" 7911d3f7af14a614617e38245fedf98a724e46a9
the behavior has changed, so that 1 now means only allow access
to processes with active perf events, with 2 indicating the old
allow-anyone-access behavior.
.TP
.IR /sys/bus/event_source/devices/*/format/ " (since Linux 3.4)"
.\" commit 641cc938815dfd09f8fa1ec72deb814f0938ac33
This subdirectory contains information on the architecture-specific
subfields available for programming the various
.I config
fields in the
.I 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
.IR perf_event_attr::config1 .
.TP
.IR /sys/bus/event_source/devices/*/events/ " (since Linux 3.4)"
.\" commit 641cc938815dfd09f8fa1ec72deb814f0938ac33
This subdirectory contains files with predefined 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, then 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)"
.\" commit 314d9f63f385096580e9e2a06eaa0745d92fe4ac
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
The errors returned by
.BR perf_event_open ()
can be inconsistent, and may
vary across processor architectures and performance monitoring units.
.TP
.B E2BIG
Returned if the
.I perf_event_attr
.I size
value is too small
(smaller than
.BR PERF_ATTR_SIZE_VER0 ),
too big (larger than the page size),
or larger than the kernel supports and the extra bytes are not zero.
When
.B E2BIG
is returned, the
.I perf_event_attr
.I size
field is overwritten by the kernel to be the size of the structure
it was expecting.
.TP
.B EACCES
Returned when the requested event requires
.B CAP_SYS_ADMIN
permissions (or a more permissive perf_event paranoid setting).
Some common cases where an unprivileged process
may encounter this error:
attaching to a process owned by a different user;
monitoring all processes on a given CPU (i.e., specifying the
.I pid
argument as \-1);
and not setting
.I exclude_kernel
when the paranoid setting requires it.
.TP
.B EBADF
Returned if the
.I group_fd
file descriptor is not valid, or, if
.B PERF_FLAG_PID_CGROUP
is set,
the cgroup file descriptor in
.I pid
is not valid.
.TP
.BR EBUSY " (since Linux 4.1)"
.\" bed5b25ad9c8a2f5d735ef0bc746ec870c01c1b0
Returned if another event already has exclusive
access to the PMU.
.TP
.B EFAULT
Returned if the
.I attr
pointer points at an invalid memory address.
.TP
.B EINVAL
Returned if the specified event is invalid.
There are many possible reasons for this.
A not-exhaustive list:
.I sample_freq
is higher than the maximum setting;
the
.I cpu
to monitor does not exist;
.I read_format
is out of range;
.I sample_type
is out of range;
the
.I flags
value is out of range;
.I exclusive
or
.I pinned
set and the event is not a group leader;
the event
.I config
values are out of range or set reserved bits;
the generic event selected is not supported; or
there is not enough room to add the selected event.
.TP
.B EMFILE
Each opened event uses one file descriptor.
If a large number of events are opened,
the per-process limit on the number of open file descriptors will be reached,
and no more events can be created.
.TP
.B ENODEV
Returned when the event involves a feature not supported
by the current CPU.
.TP
.B ENOENT
Returned if the
.I type
setting is not valid.
This error is also returned for
some unsupported generic events.
.TP
.B ENOSPC
Prior to Linux 3.3, if there was not enough room for the event,
.\" commit aa2bc1ade59003a379ffc485d6da2d92ea3370a6
.B ENOSPC
was returned.
In Linux 3.3, this was changed to
.BR EINVAL .
.B ENOSPC
is still returned if you try to add more breakpoint events
than supported by the hardware.
.TP
.B ENOSYS
Returned if
.B PERF_SAMPLE_STACK_USER
is set in
.I sample_type
and it is not supported by hardware.
.TP
.B EOPNOTSUPP
Returned if an event requiring a specific hardware feature is
requested but there is no hardware support.
This includes requesting low-skid events if not supported,
branch tracing if it is not available, sampling if no PMU
interrupt is available, and branch stacks for software events.
.TP
.B EPERM
Returned on many (but not all) architectures when an unsupported
.IR exclude_hv ", " exclude_idle ", " exclude_user ", or " exclude_kernel
setting is specified.

It can also happen, as with
.BR EACCES ,
when the requested event requires
.B CAP_SYS_ADMIN
permissions (or a more permissive perf_event paranoid setting).
This includes setting a breakpoint on a kernel address,
and (since Linux 3.13) setting a kernel function-trace tracepoint.
.\" commit a4e95fc2cbb31d70a65beffeaf8773f881328c34
.TP
.B ESRCH
Returned if attempting to attach to a process that does not exist.
.SH VERSION
.BR perf_event_open ()
was introduced in Linux 2.6.31 but was called
.\" commit 0793a61d4df8daeac6492dbf8d2f3e5713caae5e
.BR perf_counter_open ().
It was renamed in Linux 2.6.32.
.\" commit cdd6c482c9ff9c55475ee7392ec8f672eddb7be6
.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.
.\" commit ba0a6c9f6fceed11c6a99e8326f0477fe383e6b5

Prior to Linux 2.6.33 (at least for x86),
.\" commit b690081d4d3f6a23541493f1682835c3cd5c54a1
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.
.\" FIXME: cannot find a kernel commit for this one

Prior to Linux 2.6.34, there was a bug when multiplexing where the
wrong results could be returned.
.\" commit 45e16a6834b6af098702e5ea6c9a40de42ff77d8

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.
.\" commit 38b435b16c36b0d863efcf3f07b34a6fac9873fd

Prior to Linux 2.6.35,
.\" commit 050735b08ca8a016bbace4445fa025b88fee770b
.B PERF_FORMAT_GROUP
did not work with attached processes.

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
nonzero value in it.
.\" commit 4ec8363dfc1451f8c8f86825731fe712798ada02

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.
.\" commit 724b6daa13e100067c30cfc4d1ad06629609dc4e

From Linux 3.4 to Linux 3.11, the mmap
.\" commit fa7315871046b9a4c48627905691dbde57e51033
.I cap_usr_rdpmc
and
.I cap_usr_time
bits mapped to the same location.
Code should migrate to the new
.I cap_user_rdpmc
and
.I cap_user_time
fields instead.

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.
.\" commit f287d332ce835f77a4f5077d2c0ef1e3f9ea42d2
.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>

static 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)