NAME¶

pfm_get_event_info - get event information

SYNOPSIS¶

#include <perfmon/pfmlib.h>

int pfm_get_event_info(int idx, pfm_os_t os, pfm_event_info_t *info);

DESCRIPTION¶

This function returns in info information about a specific event designated by its opaque unique identifier in idx for the operating system specified in os.

The pfm_event_info_t structure is defined as follows:

typedef struct {


        const char              *name;


        const char              *desc;


        const char              *equiv;


        size_t			size;


        uint64_t                code;


        pfm_pmu_t               pmu;


        pfm_dtype_t             dtype


        int                     idx;


        int                     nattrs;


        struct {


          unsigned int is_precise:1;


          unsigned int reserved_bits:31;


        };
} pfm_event_info_t;

The fields of this structure are defined as follows:

name

This is the name of the event. This is a read-only string.

desc

This is the description of the event. This is a read-only string. It may contain multiple sentences.

equiv

Certain events may be just variations of actual events. They may be provided as handy shortcuts to avoid supplying a long list of attributes. For those events, this field is not NULL and contains the complete equivalent event string.

code

This is the raw event code. It should not be confused with the encoding of the event. This field represents only the event selection code, it does not include any unit mask or attribute settings.

pmu

This is the identification of the PMU model this event belongs to. It is of type pfm_pmu_t. Using this value and the pfm_get_pmu_info function, it is possible to get PMU information.

dtype

This field returns the representation of the event data. By default, it is PFM_DATA_UINT64.

idx This is the event unique opaque identifier. It is identical to the idx passed to the call and is provided for completeness.

nattrs

This is the number of attributes supported by this event. Attributes may be unit masks or modifiers. If the event has not attribute, then the value of this field is simply 0.

size

This field contains the size of the struct passed. This field is used to provide for extensibility of the struct without compromising backward compatibility. The value should be set to sizeof(pfm_event_info_t). If instead, a value of 0 is specified, the library assumes the struct passed is identical to the first ABI version which size is PFM_EVENT_INFO_ABI0. Thus, if fields were added after the first ABI, they will not be set by the library. The library does check that bytes beyond what is implemented are zeroes.

is_precise

This bitfield indicates whether or not the event support precise sampling. Precise sampling is a hardware mechanism that avoids instruction address skid when using interrupt-based sampling. When the event has umasks, this field means that at least one umask supports precise sampling. On Intel X86 processors, this indicates whether the event supports Precise Event-Based Sampling (PEBS).

The pfm_os_t enumeration provides the following choices:

PFM_OS_NONE

The returned information pertains only to what the PMU hardware exports. No operating system attributes is taken into account.

PFM_OS_PERF_EVENT

The returned information includes the actual PMU hardware and the additional attributes exported by the perf_events kernel interface. The perf_event attributes pertain only the PMU hardware. In case perf_events is not detected, an error is returned.

PFM_OS_PERF_EVENT_EXT

The returned information includes all of what is already provided by PFM_OS_PERF_EVENT plus all the software attributes controlled by perf_events, such as sampling period, precise sampling.

RETURN¶

If successful, the function returns PFM_SUCCESS and event information in info, otherwise it returns an error code.

ERRORS¶

PFMLIB_ERR_NOINIT

Library has not been initialized properly.

PFMLIB_ERR_INVAL

The idx argument is invalid or info is NULL or size is not zero.

PFMLIB_ERR_NOTSUPP

The requested os is not detected or supported.

AUTHOR¶

Stephane Eranian <eranian@gmail.com>