Section: Linux Programmer's Manual (2)
Updated: 2015-01-10
Index Return to Main Contents


perf_event_open - set up performance monitoring  


#include <linux/perf_event.h>
#include <linux/hw_breakpoint.h>

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

Note: There is no glibc wrapper for this system call; see NOTES.  


Given a list of parameters, perf_event_open() returns a file descriptor, for use in subsequent system calls (read(2), mmap(2), prctl(2), fcntl(2), etc.).

A call to 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.

Events can be enabled and disabled in two ways: via ioctl(2) and via prctl(2). When an event is disabled it does not count or generate overflows but does continue to exist and maintain its count value.

Events come in two flavors: counting and sampled. A 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 read(2) call. A sampling event periodically writes measurements to a buffer that can then be accessed via mmap(2).  


The pid and cpu arguments allow specifying which process and CPU to monitor:
pid == 0 and cpu == -1
This measures the calling process/thread on any CPU.
pid == 0 and cpu >= 0
This measures the calling process/thread only when running on the specified CPU.
pid > 0 and cpu == -1
This measures the specified process/thread on any CPU.
pid > 0 and cpu >= 0
This measures the specified process/thread only when running on the specified CPU.
pid == -1 and cpu >= 0
This measures all processes/threads on the specified CPU. This requires CAP_SYS_ADMIN capability or a /proc/sys/kernel/perf_event_paranoid value of less than 1.
pid == -1 and cpu == -1
This setting is invalid and will return an error. The 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 group_fd = -1. The rest of the group members are created with subsequent perf_event_open() calls with group_fd being set to the file descriptor of the group leader. (A single event on its own is created with 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), and so on---with each other, since they have counted events for the same set of executed instructions. The flags argument is formed by ORing together zero or more of the following values:
PERF_FLAG_FD_CLOEXEC (since Linux 3.14)
This flag enables the close-on-exec flag for the created event file descriptor, so that the file descriptor is automatically closed on execve(2). Setting the close-on-exec flags at creation time, rather than later with fcntl(2), avoids potential race conditions where the calling thread invokes perf_event_open() and fcntl(2) at the same time as another thread calls fork(2) then execve(2).
This flag tells the event to ignore the group_fd parameter except for the purpose of setting up output redirection using the PERF_FLAG_FD_OUTPUT flag.
PERF_FLAG_FD_OUTPUT (broken since Linux 2.6.35)
This flag re-routes the event's sampled output to instead be included in the mmap buffer of the event specified by group_fd.
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-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 test, then a file descriptor opened on /dev/cgroup/test (assuming cgroupfs is mounted on /dev/cgroup) must be passed as the pid parameter. cgroup monitoring is available only for system-wide events and may therefore require extra permissions. The perf_event_attr structure provides detailed configuration information for the event being created.

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 */
          __reserved_1   : 39;

    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 */


The fields of the perf_event_attr structure are described in more detail below:

This field specifies the overall event type. It has one of the following values:
This indicates one of the "generalized" hardware events provided by the kernel. See the config field definition for more details.
This indicates one of the software-defined events provided by the kernel (even if no hardware support is available).
This indicates a tracepoint provided by the kernel tracepoint infrastructure.
This indicates a hardware cache event. This has a special encoding, described in the config field definition.
This indicates a "raw" implementation-specific event in the config field.
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.
dynamic PMU
Since Linux 2.6.39, perf_event_open() can support multiple PMUs. To enable this, a value exported by the kernel can be used in the 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 /sys/bus/event_source/devices. In each subdirectory there is a type file whose content is an integer that can be used in the type field. For instance, /sys/bus/event_source/devices/cpu/type contains the value for the core CPU PMU, which is usually 4.
The size of the perf_event_attr structure for forward/backward compatibility. Set this using sizeof(struct perf_event_attr) to allow the kernel to see the struct size at the time of compilation.

The related define PERF_ATTR_SIZE_VER0 is set to 64; this was the size of the first published struct. PERF_ATTR_SIZE_VER1 is 72, corresponding to the addition of breakpoints in Linux 2.6.33. PERF_ATTR_SIZE_VER2 is 80 corresponding to the addition of branch sampling in Linux 3.4. PERF_ATTR_SIZE_VER3 is 96 corresponding to the addition of sample_regs_user and sample_stack_user in Linux 3.7.

This specifies which event you want, in conjunction with the type field. The 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 config field that are dependent on the value of the previously described type field. What follows are various possible settings for config separated out by type.

If type is PERF_TYPE_HARDWARE, we are measuring one of the generalized hardware CPU events. Not all of these are available on all platforms. Set config to one of the following:

Total cycles. Be wary of what happens during CPU frequency scaling.
Retired instructions. Be careful, these can be affected by various issues, most notably hardware interrupt counts.
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.
Cache misses. Usually this indicates Last Level Cache misses; this is intended to be used in conjunction with the PERF_COUNT_HW_CACHE_REFERENCES event to calculate cache miss rates.
Retired branch instructions. Prior to Linux 2.6.34, this used the wrong event on AMD processors.
Mispredicted branch instructions.
Bus cycles, which can be different from total cycles.
Stalled cycles during issue.
Stalled cycles during retirement.
Total cycles; not affected by CPU frequency scaling.
If type is PERF_TYPE_SOFTWARE, we are measuring software events provided by the kernel. Set config to one of the following:
This reports the CPU clock, a high-resolution per-CPU timer.
This reports a clock count specific to the task that is running.
This reports the number of page faults.
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.
This reports the number of times the process has migrated to a new CPU.
This counts the number of minor page faults. These did not require disk I/O to handle.
This counts the number of major page faults. These required disk I/O to handle.
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).
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.
PERF_COUNT_SW_DUMMY (since Linux 3.12)
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.

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

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

    (perf_hw_cache_id) | (perf_hw_cache_op_id << 8) |
    (perf_hw_cache_op_result_id << 16)
where perf_hw_cache_id is one of:
for measuring Level 1 Data Cache
for measuring Level 1 Instruction Cache
for measuring Last-Level Cache
for measuring the Data TLB
for measuring the Instruction TLB
for measuring the branch prediction unit
PERF_COUNT_HW_CACHE_NODE (since Linux 3.1)
for measuring local memory accesses
and perf_hw_cache_op_id is one of
for read accesses
for write accesses
for prefetch accesses
and perf_hw_cache_op_result_id is one of
to measure accesses
to measure misses

If type is PERF_TYPE_RAW, then a custom "raw" 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 perf_event_open() expects in this field.

If type is PERF_TYPE_BREAKPOINT, then leave config set to zero. Its parameters are set in other places.

sample_period, sample_freq
A "sampling" event is one that generates an overflow notification every N events, where N is given by sample_period. A sampling event has sample_period > 0. When an overflow occurs, requested data is recorded in the mmap buffer. The sample_type field controls what data is recorded on each overflow.

sample_freq can be used if you wish to use frequency rather than period. In this case, you set the freq flag. The kernel will adjust the sampling period to try and achieve the desired rate. The rate of adjustment is a timer tick.

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 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 enum perf_event_sample_format order.
Records instruction pointer.
Records the process and thread IDs.
Records a timestamp.
Records an address, if applicable.
Record counter values for all events in a group, not just the group leader.
Records the callchain (stack backtrace).
Records a unique ID for the opened event's group leader.
Records CPU number.
Records the current sampling period.
Records a unique ID for the opened event. Unlike 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.
Records additional data, if applicable. Usually returned by tracepoint events.
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 branch_sample_type field for how to filter which branches are reported.

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).
PERF_SAMPLE_STACK_USER (since Linux 3.7)
Records the user level stack, allowing stack unwinding.
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.
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.
Places the 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 sample_type settings. Parsing the event stream properly was not possible because the format of the record was needed to find SAMPLE_ID, but the format could not be found without knowing what event the sample belonged to (causing a circular dependency).

The PERF_SAMPLE_IDENTIFIER setting makes the event stream always parsable by putting SAMPLE_ID in a fixed location, even though it means having duplicate SAMPLE_ID values in records.

Records reasons for transactional memory abort events (for example, from Intel TSX transactional memory support).

The 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).

This field specifies the format of the data returned by read(2) on a perf_event_open() file descriptor.
Adds the 64-bit time_enabled field. This can be used to calculate estimated totals if the PMU is overcommitted and multiplexing is happening.
Adds the 64-bit time_running field. This can be used to calculate estimated totals if the PMU is overcommitted and multiplexing is happening.
Adds a 64-bit unique value that corresponds to the event group.
Allows all counter values in an event group to be read with one read.
The disabled bit specifies whether the counter starts out disabled or enabled. If disabled, the event can later be enabled by ioctl(2), prctl(2), or enable_on_exec.

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

The 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 read_formats, such as PERF_FORMAT_GROUP.

The 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., read(2) returns 0) until the counter is subsequently enabled or disabled.
The 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 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).

If this bit is set, the count excludes events that happen in user space.
If this bit is set, the count excludes events that happen in kernel-space.
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.
If set, don't count when the CPU is idle.
The mmap bit enables generation of PERF_RECORD_MMAP samples for every mmap(2) call that has 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.
The comm bit enables tracking of process command name as modified by the exec(2) and prctl(PR_SET_NAME) system calls as well as writing to /proc/self/comm. If the comm_exec flag is also successfully set (possible since Linux 3.16), then the misc flag PERF_RECORD_MISC_COMM_EXEC can be used to differentiate the exec(2) case from the others.
If this bit is set, then sample_frequency not sample_period is used when setting up the sampling interval.
This bit enables saving of event counts on context switch for inherited tasks. This is meaningful only if the inherit field is set.
If this bit is set, a counter is automatically enabled after a call to exec(2).
If this bit is set, then fork/exit notifications are included in the ring buffer.
If set, have an overflow notification happen when we cross the wakeup_watermark boundary. Otherwise, overflow notifications happen after wakeup_events samples.
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:

0 -
SAMPLE_IP can have arbitrary skid.
1 -
SAMPLE_IP must have constant skid.
2 -
SAMPLE_IP requested to have 0 skid.
3 -
SAMPLE_IP must have 0 skid. See also PERF_RECORD_MISC_EXACT_IP.
mmap_data (since Linux 2.6.36)
The counterpart of the mmap field. This enables generation of PERF_RECORD_MMAP samples for mmap(2) calls that do not have PROT_EXEC set (for example data and SysV shared memory).
sample_id_all (since Linux 2.6.38)
If set, then TID, TIME, ID, STREAM_ID, and CPU can additionally be included in non-PERF_RECORD_SAMPLEs if the corresponding sample_type is selected.

If 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 id value appearing twice.

The layout is described by this pseudo-structure:

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 */
exclude_host (since Linux 3.2)
Do not measure time spent in VM host.
exclude_guest (since Linux 3.2)
Do not measure time spent in VM guest.
exclude_callchain_kernel (since Linux 3.7)
Do not include kernel callchains.
exclude_callchain_user (since Linux 3.7)
Do not include user callchains.
mmap2 (since Linux 3.16)
Generate an extended executable mmap record that contains enough additional information to uniquely identify shared mappings. The mmap flag must also be set for this to work.
comm_exec (since Linux 3.16)
This is purely a feature-detection flag, it does not change kernel behavior. If this flag can successfully be set, then, when comm is enabled, the PERF_RECORD_MISC_COMM_EXEC flag will be set in the misc field of a comm record header if the rename event being reported was caused by a call to exec(2). This allows tools to distinguish between the various types of process renaming.
wakeup_events, wakeup_watermark
This union sets how many samples (wakeup_events) or bytes (wakeup_watermark) happen before an overflow notification happens. Which one is used is selected by the watermark bit flag.

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

Prior to Linux 3.0 setting wakeup_events to 0 resulted in no overflow notifications; more recent kernels treat 0 the same as 1.

bp_type (since Linux 2.6.33)
This chooses the breakpoint type. It is one of:
No breakpoint.
Count when we read the memory location.
Count when we write the memory location.
Count when we read or write the memory location.
Count when we execute code at the memory location.

The values can be combined via a bitwise or, but the combination of HW_BREAKPOINT_R or HW_BREAKPOINT_W with HW_BREAKPOINT_X is not allowed.

bp_addr (since Linux 2.6.33)
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.
config1 (since Linux 2.6.39)
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.
bp_len (since Linux 2.6.33)
bp_len is the length of the breakpoint being measured if type is PERF_TYPE_BREAKPOINT. Options are HW_BREAKPOINT_LEN_1, HW_BREAKPOINT_LEN_2, HW_BREAKPOINT_LEN_4, HW_BREAKPOINT_LEN_8. For an execution breakpoint, set this to sizeof(long).
config2 (since Linux 2.6.39)

config2 is a further extension of the config1 field.

branch_sample_type (since Linux 3.4)
If 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.

Branch target is in user space.
Branch target is in kernel space.
Branch target is in hypervisor.
A convenience value that is the three preceding values ORed together.

In addition to the privilege value, at least one or more of the following bits must be set.

Any branch type.
Any call branch.
Any return branch.
Indirect calls.
PERF_SAMPLE_BRANCH_COND (since Linux 3.16)
Conditional branches.
Transactional memory aborts.
PERF_SAMPLE_BRANCH_IN_TX (since Linux 3.11)
Branch in transactional memory transaction.
PERF_SAMPLE_BRANCH_NO_TX (since Linux 3.11)
Branch not in transactional memory transaction.

sample_regs_user (since Linux 3.7)
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 arch/ARCH/include/uapi/asm/perf_regs.h.
sample_stack_user (since Linux 3.7)
This defines the size of the user stack to dump if PERF_SAMPLE_STACK_USER is specified.

Reading results

Once a 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 read_format field in the attr structure at open time.

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

Here is the layout of the data returned by a read:

If PERF_FORMAT_GROUP was specified to allow reading all events in a group at once:

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];
If PERF_FORMAT_GROUP was not specified:

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 */

The values read are as follows:

The number of events in this file descriptor. Only available if PERF_FORMAT_GROUP was specified.
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 time_enabled and time running values can be used to scale an estimated value for the count.
An unsigned 64-bit value containing the counter result.
A globally unique value for this particular event, only there if PERF_FORMAT_ID was specified in read_format.

MMAP layout

When using perf_event_open() in sampled mode, asynchronous events (like counter overflow or PROT_EXEC mmap tracking) are logged into a ring-buffer. This ring-buffer is created and accessed through mmap(2).

The mmap size should be 1+2^n pages, where the first page is a metadata page (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:

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 */

The following list describes the fields in the perf_event_mmap_page structure in more detail:

Version number of this structure.
The lowest version this is compatible with.
A seqlock for synchronization.
A unique hardware counter identifier.
When using rdpmc for reads this offset value must be added to the one returned by rdpmc to get the current total event count.
Time the event was active.
Time the event was running.
cap_usr_time / cap_usr_rdpmc / cap_bit0 (since Linux 3.4)
There was a bug in the definition of cap_usr_time and 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 cap_usr_time or cap_usr_rdpmc were actually set.

Starting with Linux 3.12, these are renamed to cap_bit0 and you should use the cap_user_time and cap_user_rdpmc fields instead.

cap_bit0_is_deprecated (since Linux 3.12)
If set, this bit indicates that the kernel supports the properly separated cap_user_time and cap_user_rdpmc bits.

If not-set, it indicates an older kernel where cap_usr_time and cap_usr_rdpmc map to the same bit and thus both features should be used with caution.

cap_user_rdpmc (since Linux 3.12)
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:

u32 seq, time_mult, time_shift, idx, width;
u64 count, enabled, running;
u64 cyc, time_offset;

do {
    seq = pc->lock;
    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);

} while (pc->lock != seq);
cap_user_time (since Linux 3.12)
This bit indicates the hardware has a constant, nonstop timestamp counter (TSC on x86).
cap_user_time_zero (since Linux 3.12)
Indicates the presence of time_zero which allows mapping timestamp values to the hardware clock.
If 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:

pmc <<= 64 - pmc_width;
pmc >>= 64 - pmc_width; // signed shift right
count += pmc;
time_shift, time_mult, time_offset

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

    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);

Where time_offset, time_mult, time_shift, and 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:

    enabled += delta;
    if (idx)
        running += delta;
    quot = count / running;
    rem  = count % running;
    count = quot * enabled + (rem * enabled) / running;
time_zero (since Linux 3.12)

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

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

And vice versa:

    quot = cyc >> time_shift;
    rem  = cyc & ((1 << time_shift) - 1);
    timestamp = time_zero + quot * time_mult +
        ((rem * time_mult) >> time_shift);
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().

When the mapping is PROT_WRITE, the 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.

The following 2^n ring-buffer pages have the layout described below.

If 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 PERF_RECORD_SAMPLE below, it will be stashed just after the perf_event_header and the fields already present for the existing fields, that is, at the end of the payload. That way a newer file will be supported by older perf tools, with these new optional fields being ignored.

The mmap values start with a header:

struct perf_event_header {
    __u32   type;
    __u16   misc;
    __u16   size;

Below, we describe the perf_event_header fields in more detail. For ease of reading, the fields with shorter descriptions are presented first.

This indicates the size of the record.
The misc field contains additional information about the sample.

The CPU mode can be determined from this value by masking with 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):

Unknown CPU mode.
Sample happened in the kernel.
Sample happened in user code.
Sample happened in the hypervisor.
Sample happened in the guest kernel.
Sample happened in guest user code.

In addition, one of the following bits can be set:
This is set when the mapping is not executable; otherwise the mapping is executable.
This is set for a PERF_RECORD_COMM record on kernels more recent than Linux 3.16 if a process name change was caused by an exec(2) system call. It is an alias for PERF_RECORD_MISC_MMAP_DATA since the two values would not be set in the same record.
This indicates that the content of PERF_SAMPLE_IP points to the actual instruction that triggered the event. See also perf_event_attr.precise_ip.
This indicates there is extended data available (currently not used).
The type value is one of the below. The values in the corresponding record (that follows the header) depend on the type selected as shown.

The MMAP events record the PROT_EXEC mappings so that we can correlate user-space IPs to code. They have the following structure:

struct {
    struct perf_event_header header;
    u32    pid, tid;
    u64    addr;
    u64    len;
    u64    pgoff;
    char   filename[];
is the process ID.
is the thread ID.
is the address of the allocated memory. len is the length of the allocated memory. pgoff is the page offset of the allocated memory. filename is a string describing the backing of the allocated memory.
This record indicates when events are lost.

struct {
    struct perf_event_header header;
    u64 id;
    u64 lost;
    struct sample_id sample_id;
is the unique event ID for the samples that were lost.
is the number of events that were lost.
This record indicates a change in the process name.

struct {
    struct perf_event_header header;
    u32 pid;
    u32 tid;
    char comm[];
    struct sample_id sample_id;
is the process ID.
is the thread ID.
is a string containing the new name of the process.
This record indicates a process exit event.

struct {
    struct perf_event_header header;
    u32 pid, ppid;
    u32 tid, ptid;
    u64 time;
    struct sample_id sample_id;
This record indicates a throttle/unthrottle event.

struct {
    struct perf_event_header header;
    u64 time;
    u64 id;
    u64 stream_id;
    struct sample_id sample_id;
This record indicates a fork event.

struct {
    struct perf_event_header header;
    u32 pid, ppid;
    u32 tid, ptid;
    u64 time;
    struct sample_id sample_id;
This record indicates a read event.

struct {
    struct perf_event_header header;
    u32 pid, tid;
    struct read_format values;
    struct sample_id sample_id;
This record indicates a sample.

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 */
    u64   weight;     /* if PERF_SAMPLE_WEIGHT */
    u64   data_src;   /* if PERF_SAMPLE_DATA_SRC */
    u64   transaction;/* if PERF_SAMPLE_TRANSACTION */
If PERF_SAMPLE_IDENTIFIER is enabled, a 64-bit unique ID is included. This is a duplication of the PERF_SAMPLE_ID id value, but included at the beginning of the sample so parsers can easily obtain the value.
If PERF_SAMPLE_IP is enabled, then a 64-bit instruction pointer value is included.
pid, tid
If PERF_SAMPLE_TID is enabled, then a 32-bit process ID and 32-bit thread ID are included.
If 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.
If 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.
If 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 PERF_FORMAT_ID.
If PERF_SAMPLE_STREAM_ID is enabled, a 64-bit unique ID is included. Unlike 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.
cpu, res
If 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.
If PERF_SAMPLE_PERIOD is enabled, a 64-bit value indicating the current sampling period is written.
If 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 read_format value used at perf_event_open() time.
nr, ips[nr]
If 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.
size, data[size]
If 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.

bnr, lbr[bnr]
If PERF_SAMPLE_BRANCH_STACK is enabled, then a 64-bit value indicating the number of records is included, followed by bnr perf_branch_entry structures which each include the fields:
This indicates the source instruction (may not be a branch).
The branch target.
The branch target was mispredicted.
The branch target was predicted.
in_tx (since Linux 3.11)
The branch was in a transactional memory transaction.
abort (since Linux 3.11)
The branch was in an aborted transactional memory transaction.

The entries are from most to least recent, so the first entry has the most recent branch.

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

The type of branches recorded is specified by the branch_sample_type field.

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


The regs field is an array of the CPU registers that were specified by the sample_regs_user attr field. The number of values is the number of bits set in the sample_regs_user bit mask.

size, data[size], dyn_size
If PERF_SAMPLE_STACK_USER is enabled, then the user stack is recorded. This can be used to generate stack backtraces. size is the size requested by the user in sample_stack_user or else the maximum record size. data is the stack data (a raw dump of the memory pointed to by the stack pointer at the time of sampling). dyn_size is the amount of data actually dumped (can be less than size).
If 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.
If PERF_SAMPLE_DATA_SRC is enabled, then a 64-bit value is recorded that is made up of the following fields:
Type of opcode, a bitwise combination of:

Not available
Load instruction
Store instruction
Executable code
Memory hierarchy level hit or miss, a bitwise combination of the following, shifted left by PERF_MEM_LVL_SHIFT:

Not available
Level 1 cache
Line fill buffer
Level 2 cache
Level 3 cache
Local DRAM
Remote DRAM 1 hop
Remote DRAM 2 hops
Remote cache 1 hop
Remote cache 2 hops
I/O memory
Uncached memory
Snoop mode, a bitwise combination of the following, shifted left by PERF_MEM_SNOOP_SHIFT:

Not available
No snoop
Snoop hit
Snoop miss
Snoop hit modified
Lock instruction, a bitwise combination of the following, shifted left by PERF_MEM_LOCK_SHIFT:

Not available
Locked transaction
TLB access hit or miss, a bitwise combination of the following, shifted left by PERF_MEM_TLB_SHIFT:

Not available
Level 1 TLB
Level 2 TLB
Hardware walker
OS fault handler
If the 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:

Abort from an elision type transaction (Intel-CPU-specific).
Abort from a generic transaction.
Synchronous abort (related to the reported instruction).
Asynchronous abort (not related to the reported instruction).
Retryable abort (retrying the transaction may have succeeded).
Abort due to memory conflicts with other threads.
Abort due to write capacity overflow.
Abort due to read capacity overflow.
In addition, a user-specified abort code can be obtained from the high 32 bits of the field by shifting right by PERF_TXN_ABORT_SHIFT and masking with PERF_TXN_ABORT_MASK.
This record includes extended information on mmap(2) calls returning executable mappings. The format is similar to that of the PERF_RECORD_MMAP record, but includes extra values that allow uniquely identifying shared mappings.

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;
is the process ID.
is the thread ID.
is the address of the allocated memory.
is the length of the allocated memory.
is the page offset of the allocated memory.
is the major ID of the underlying device.
is the minor ID of the underlying device.
is the inode number.
is the inode generation.
is the protection information.
is the flags information.
is a string describing the backing of the allocated memory.

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 poll(2), select(2), or epoll(2). Alternately, a SIGIO signal handler can be created and the event configured with fcntl(2) to generate SIGIO signals.

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

There are two ways to generate overflow notifications.

The first is to set a wakeup_events or 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 POLL_IN is indicated.

The other way is by use of the PERF_EVENT_IOC_REFRESH ioctl. This ioctl adds to a counter that decrements each time the event overflows. When nonzero, POLL_IN is indicated, but once the counter reaches 0 POLL_HUP is indicated and the underlying event is disabled.

Starting with Linux 3.18, POLL_HUP is indicated if the event being monitored is attached to a different process and that process exits.  

rdpmc instruction

Starting with Linux 3.4 on x86, you can use the rdpmc instruction to get low-latency reads without having to enter the kernel. Note that using rdpmc is not necessarily faster than other methods for reading event values.

Support for this can be detected with the cap_usr_rdpmc field in the mmap page; documentation on how to calculate event values can be found in that section.  

perf_event ioctl calls

Various ioctls act on perf_event_open() file descriptors:

This enables the individual event or event group specified by the file descriptor argument.

If the 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).

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 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).

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 POLL_IN set will happen on each overflow until the count reaches 0; when that happens a notification with POLL_HUP set is sent and the event is disabled. Using an argument of 0 is considered undefined behavior.
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 time_enabled or time_running values.

If the 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).

This updates the overflow period for the event.

Since Linux 3.7 (on ARM) and Linux 3.14 (all other architectures), 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 this ioctl always failed due to a bug in the kernel.

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.

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.

PERF_EVENT_IOC_ID (since Linux 3.12)
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.


Using prctl

A process can enable or disable all the event groups that are attached to it using the prctl(2) PR_TASK_PERF_EVENTS_ENABLE and 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.  

perf_event related configuration files

Files in /proc/sys/kernel/

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

only allow user-space measurements.
allow both kernel and user measurements (default).
allow access to CPU-specific data but not raw tracepoint samples.
no restrictions.
The existence of the perf_event_paranoid file is the official method for determining if a kernel supports perf_event_open().

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).


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

Files in /sys/bus/event_source/devices/
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 /sys/bus/event_source/devices/. Each subdirectory corresponds to a different PMU.
/sys/bus/event_source/devices/*/type (since Linux 2.6.38)
This contains an integer that can be used in the type field of perf_event_attr to indicate that you wish to use this PMU.
/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.
/sys/bus/event_source/devices/*/format/ (since Linux 3.4)
This subdirectory contains information on the architecture-specific subfields available for programming the various 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 event may contain the value 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.

/sys/bus/event_source/devices/*/events/ (since Linux 3.4)
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 ./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: event=0x2,inv,ldlat=3.

This file is the standard kernel device interface for injecting hotplug events.
/sys/bus/event_source/devices/*/cpumask (since Linux 3.7)
The 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.


perf_event_open() returns the new file descriptor, or -1 if an error occurred (in which case, errno is set appropriately).  


The errors returned by perf_event_open() can be inconsistent, and may vary across processor architectures and performance monitoring units.
Returned if the perf_event_attr size value is too small (smaller than 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 E2BIG is returned, the perf_event_attr size field is overwritten by the kernel to be the size of the structure it was expecting.
Returned when the requested event requires 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 pid argument as -1); and not setting exclude_kernel when the paranoid setting requires it.
Returned if the group_fd file descriptor is not valid, or, if PERF_FLAG_PID_CGROUP is set, the cgroup file descriptor in pid is not valid.
Returned if the attr pointer points at an invalid memory address.
Returned if the specified event is invalid. There are many possible reasons for this. A not-exhaustive list: sample_freq is higher than the maximum setting; the cpu to monitor does not exist; read_format is out of range; sample_type is out of range; the flags value is out of range; exclusive or pinned set and the event is not a group leader; the event 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.
Each opened event uses one file descriptor. If a large number of events are opened the per-user file descriptor limit (often 1024) will be hit and no more events can be created.
Returned when the event involves a feature not supported by the current CPU.
Returned if the type setting is not valid. This error is also returned for some unsupported generic events.
Prior to Linux 3.3, if there was not enough room for the event, ENOSPC was returned. In Linux 3.3, this was changed to EINVAL. ENOSPC is still returned if you try to add more breakpoint events than supported by the hardware.
Returned if PERF_SAMPLE_STACK_USER is set in sample_type and it is not supported by hardware.
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.
Returned on many (but not all) architectures when an unsupported exclude_hv, exclude_idle, exclude_user, or exclude_kernel setting is specified.

It can also happen, as with EACCES, when the requested event requires 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.

Returned if attempting to attach to a process that does not exist.


perf_event_open() was introduced in Linux 2.6.31 but was called perf_counter_open(). It was renamed in Linux 2.6.32.  


This perf_event_open() system call Linux- specific and should not be used in programs intended to be portable.  


Glibc does not provide a wrapper for this system call; call it using syscall(2). See the example below.

The official way of knowing if perf_event_open() support is enabled is checking for the existence of the file /proc/sys/kernel/perf_event_paranoid.  


The F_SETOWN_EX option to 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 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, 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 nonzero value in it.

From Linux 2.6.31 to Linux 3.4, the 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.

From Linux 3.4 to Linux 3.11, the mmap cap_usr_rdpmc and cap_usr_time bits mapped to the same location. Code should migrate to the new cap_user_rdpmc and 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.  


The following is a short example that measures the total instruction count of a call to printf(3).

#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;

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.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);

    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);



fcntl(2), mmap(2), open(2), prctl(2), read(2)



Reading results
MMAP layout
Overflow handling
rdpmc instruction
perf_event ioctl calls
Using prctl
perf_event related configuration files

This document was created by man2html, using the manual pages.
Time: 15:57:00 GMT, January 23, 2015