LDP
/
man-pages
mirror of https://github.com/mkerrisk/man-pages
0
1
Fork 0
Code Releases Activity

perf_event_open.2: srcfix 

Signed-off-by: Alejandro Colomar <alx.manpages@gmail.com>
Signed-off-by: Michael Kerrisk <mtk.manpages@gmail.com>
This commit is contained in:
Alejandro Colomar 2020-11-13 21:34:56 +01:00 committed by Michael Kerrisk
parent 17cf67cc72
commit 2b9bf369d1
1 changed files with 122 additions and 93 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

215
man2/perf_event_open.2
View File

@ -126,7 +126,7 @@ The leader is created first, with
The rest of the group members are created with subsequent The rest of the group members are created with subsequent
.BR perf_event_open () .BR perf_event_open ()
calls with calls with
.IR group_fd .I group_fd
being set to the file descriptor of the group leader. being set to the file descriptor of the group leader.
(A single event on its own is created with (A single event on its own is created with
.IR group_fd " = \-1" .IR group_fd " = \-1"
@ -162,7 +162,7 @@ then
.TP .TP
.BR PERF_FLAG_FD_NO_GROUP .BR PERF_FLAG_FD_NO_GROUP
This flag tells the event to ignore the This flag tells the event to ignore the
.IR group_fd .I group_fd
parameter except for the purpose of setting up output redirection parameter except for the purpose of setting up output redirection
using the using the
.B PERF_FLAG_FD_OUTPUT .B PERF_FLAG_FD_OUTPUT
@ -352,7 +352,11 @@ These two dynamic PMUs create a kprobe/uprobe and attach it to the
file descriptor generated by perf_event_open. file descriptor generated by perf_event_open.
The kprobe/uprobe will be destroyed on the destruction of the file descriptor. The kprobe/uprobe will be destroyed on the destruction of the file descriptor.
See fields See fields
.IR kprobe_func ", " uprobe_path ", " kprobe_addr ", and " probe_offset .IR kprobe_func ,
.IR uprobe_path ,
.IR kprobe_addr ,
and
.I probe_offset
for more details. for more details.
.RE .RE
.TP .TP
@ -401,7 +405,9 @@ the
.I type .I type
field. field.
The The
.IR config1 " and " config2 .I config1
and
.I config2
fields are also taken into account in cases where 64 bits is not fields are also taken into account in cases where 64 bits is not
enough to fully specify the event. enough to fully specify the event.
The encoding of these fields are event dependent. The encoding of these fields are event dependent.
@ -558,15 +564,15 @@ then we are measuring a hardware CPU cache event.
To calculate the appropriate To calculate the appropriate
.I config .I config
value use the following equation: value use the following equation:
.RS 4
.PP .PP
.RS 4 .in +4n
.RS 4
.EX .EX
config = (perf_hw_cache_id) | config = (perf_hw_cache_id) |
(perf_hw_cache_op_id << 8) | (perf_hw_cache_op_id << 8) |
(perf_hw_cache_op_result_id << 16); (perf_hw_cache_op_result_id << 16);
.EE .EE
.RE .in
.PP .PP
where where
.I perf_hw_cache_id .I perf_hw_cache_id
@ -652,24 +658,28 @@ Its parameters are set in other places.
If If
.I type .I type
is is
.BR kprobe .B kprobe
or or
.BR uprobe , .BR uprobe ,
set set
.IR retprobe .I retprobe
(bit 0 of (bit 0 of
.IR config , .IR config ,
see see
.IR /sys/bus/event_source/devices/[k,u]probe/format/retprobe ) .IR /sys/bus/event_source/devices/[k,u]probe/format/retprobe )
for kretprobe/uretprobe. for kretprobe/uretprobe.
See fields See fields
.IR kprobe_func ", " uprobe_path ", " kprobe_addr ", and " probe_offset .IR kprobe_func ,
.IR uprobe_path ,
.IR kprobe_addr ,
and
.I probe_offset
for more details. for more details.
.RE .RE
.TP .TP
.IR kprobe_func ", " uprobe_path ", " kprobe_addr ", and " probe_offset .IR kprobe_func ", " uprobe_path ", " kprobe_addr ", and " probe_offset
These fields describe the kprobe/uprobe for dynamic PMUs These fields describe the kprobe/uprobe for dynamic PMUs
.BR kprobe .B kprobe
and and
.BR uprobe . .BR uprobe .
For For
@ -712,7 +722,7 @@ to try and achieve the desired rate.
The rate of adjustment is a The rate of adjustment is a
timer tick. timer tick.
.TP .TP
.I "sample_type" .I sample_type
The various bits in this field specify which values to include The various bits in this field specify which values to include
in the sample. in the sample.
They will be recorded in a ring-buffer, They will be recorded in a ring-buffer,
@ -859,7 +869,7 @@ hardware at the time of the sampled
instruction's retirement. instruction's retirement.
.RE .RE
.TP .TP
.IR "read_format" .I read_format
This field specifies the format of the data returned by This field specifies the format of the data returned by
.BR read (2) .BR read (2)
on a on a
@ -888,7 +898,7 @@ 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. Allows all counter values in an event group to be read with one read.
.RE .RE
.TP .TP
.IR "disabled" .I disabled
The The
.I disabled .I disabled
bit specifies whether the counter starts out disabled or enabled. bit specifies whether the counter starts out disabled or enabled.
@ -909,7 +919,7 @@ Despite
being 0, the child events will not start until the group leader being 0, the child events will not start until the group leader
is enabled. is enabled.
.TP .TP
.IR "inherit" .I inherit
The The
.I inherit .I inherit
bit specifies that this counter should count events of child bit specifies that this counter should count events of child
@ -923,7 +933,7 @@ Inherit does not work for some combinations of
values, such as values, such as
.BR PERF_FORMAT_GROUP . .BR PERF_FORMAT_GROUP .
.TP .TP
.IR "pinned" .I pinned
The The
.I pinned .I pinned
bit specifies that the counter should always be on the CPU if at all bit specifies that the counter should always be on the CPU if at all
@ -936,7 +946,7 @@ return end-of-file (i.e.,
.BR read (2) .BR read (2)
returns 0) until the counter is subsequently enabled or disabled. returns 0) until the counter is subsequently enabled or disabled.
.TP .TP
.IR "exclusive" .I exclusive
The The
.I exclusive .I exclusive
bit specifies that when this counter's group is on the CPU, bit specifies that when this counter's group is on the CPU,
@ -952,13 +962,13 @@ This includes any users running a system-wide
measurement as well as any kernel use of the performance counters measurement as well as any kernel use of the performance counters
(including the commonly enabled NMI Watchdog Timer interface). (including the commonly enabled NMI Watchdog Timer interface).
.TP .TP
.IR "exclude_user" .I exclude_user
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 user space.
.TP .TP
.IR "exclude_kernel" .I exclude_kernel
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 kernel space.
.TP .TP
.IR "exclude_hv" .I exclude_hv
If this bit is set, the count excludes events that happen in the If this bit is set, the count excludes events that happen in the
hypervisor. hypervisor.
This is mainly for PMUs that have built-in support for handling this This is mainly for PMUs that have built-in support for handling this
@ -966,12 +976,12 @@ This is mainly for PMUs that have built-in support for handling this
Extra support is needed for handling hypervisor measurements on most Extra support is needed for handling hypervisor measurements on most
machines. machines.
.TP .TP
.IR "exclude_idle" .I exclude_idle
If set, don't count when the CPU is running the idle task. If set, don't count when the CPU is running the idle task.
While you can currently enable this for any event type, it is ignored While you can currently enable this for any event type, it is ignored
for all but software events. for all but software events.
.TP .TP
.IR "mmap" .I mmap
The The
.I mmap .I mmap
bit enables generation of bit enables generation of
@ -985,7 +995,7 @@ This allows tools to notice new executable code being mapped into
a program (dynamic shared libraries for example) a program (dynamic shared libraries for example)
so that addresses can be mapped back to the original code. so that addresses can be mapped back to the original code.
.TP .TP
.IR "comm" .I comm
The The
.I comm .I comm
bit enables tracking of process command name as modified by the bit enables tracking of process command name as modified by the
@ -1004,30 +1014,30 @@ can be used to differentiate the
.BR exec (2) .BR exec (2)
case from the others. case from the others.
.TP .TP
.IR "freq" .I freq
If this bit is set, then If this bit is set, then
.I sample_frequency .I sample_frequency
not not
.I sample_period .I sample_period
is used when setting up the sampling interval. is used when setting up the sampling interval.
.TP .TP
.IR "inherit_stat" .I inherit_stat
This bit enables saving of event counts on context switch for This bit enables saving of event counts on context switch for
inherited tasks. inherited tasks.
This is meaningful only if the This is meaningful only if the
.I inherit .I inherit
field is set. field is set.
.TP .TP
.IR "enable_on_exec" .I enable_on_exec
If this bit is set, a counter is automatically If this bit is set, a counter is automatically
enabled after a call to enabled after a call to
.BR exec (2). .BR exec (2).
.TP .TP
.IR "task" .I task
If this bit is set, then If this bit is set, then
fork/exit notifications are included in the ring buffer. fork/exit notifications are included in the ring buffer.
.TP .TP
.IR "watermark" .I watermark
If set, have an overflow notification happen when we cross the If set, have an overflow notification happen when we cross the
.I wakeup_watermark .I wakeup_watermark
boundary. boundary.
@ -1035,7 +1045,7 @@ Otherwise, overflow notifications happen after
.I wakeup_events .I wakeup_events
samples. samples.
.TP .TP
.IR "precise_ip" " (since Linux 2.6.35)" .IR precise_ip " (since Linux 2.6.35)"
.\" commit ab608344bcbde4f55ec4cd911b686b0ce3eae076 .\" commit ab608344bcbde4f55ec4cd911b686b0ce3eae076
This controls the amount of skid. This controls the amount of skid.
Skid is how many instructions Skid is how many instructions
@ -1064,7 +1074,7 @@ See also the description of
.BR PERF_RECORD_MISC_EXACT_IP . .BR PERF_RECORD_MISC_EXACT_IP .
.RE .RE
.TP .TP
.IR "mmap_data" " (since Linux 2.6.36)" .IR mmap_data " (since Linux 2.6.36)"
.\" commit 3af9e859281bda7eb7c20b51879cf43aa788ac2e .\" commit 3af9e859281bda7eb7c20b51879cf43aa788ac2e
This is the counterpart of the This is the counterpart of the
.I mmap .I mmap
@ -1077,7 +1087,7 @@ calls that do not have
.B PROT_EXEC .B PROT_EXEC
set (for example data and SysV shared memory). set (for example data and SysV shared memory).
.TP .TP
.IR "sample_id_all" " (since Linux 2.6.38)" .IR sample_id_all " (since Linux 2.6.38)"
.\" commit c980d1091810df13f21aabbce545fd98f545bbf7 .\" commit c980d1091810df13f21aabbce545fd98f545bbf7
If set, then TID, TIME, ID, STREAM_ID, and CPU can If set, then TID, TIME, ID, STREAM_ID, and CPU can
additionally be included in additionally be included in
@ -1109,7 +1119,7 @@ struct sample_id {
.EE .EE
.in .in
.TP .TP
.IR "exclude_host" " (since Linux 3.2)" .IR exclude_host " (since Linux 3.2)"
.\" commit a240f76165e6255384d4bdb8139895fac7988799 .\" commit a240f76165e6255384d4bdb8139895fac7988799
When conducting measurements that include processes running When conducting measurements that include processes running
VM instances (i.e., have executed a VM instances (i.e., have executed a
@ -1120,7 +1130,7 @@ This is only meaningful outside the guests; this setting does
not change counts gathered inside of a guest. not change counts gathered inside of a guest.
Currently, this functionality is x86 only. Currently, this functionality is x86 only.
.TP .TP
.IR "exclude_guest" " (since Linux 3.2)" .IR exclude_guest " (since Linux 3.2)"
.\" commit a240f76165e6255384d4bdb8139895fac7988799 .\" commit a240f76165e6255384d4bdb8139895fac7988799
When conducting measurements that include processes running When conducting measurements that include processes running
VM instances (i.e., have executed a VM instances (i.e., have executed a
@ -1131,15 +1141,15 @@ This is only meaningful outside the guests; this setting does
not change counts gathered inside of a guest. not change counts gathered inside of a guest.
Currently, this functionality is x86 only. Currently, this functionality is x86 only.
.TP .TP
.IR "exclude_callchain_kernel" " (since Linux 3.7)" .IR exclude_callchain_kernel " (since Linux 3.7)"
.\" commit d077526485d5c9b12fe85d0b2b3b7041e6bc5f91 .\" commit d077526485d5c9b12fe85d0b2b3b7041e6bc5f91
Do not include kernel callchains. Do not include kernel callchains.
.TP .TP
.IR "exclude_callchain_user" " (since Linux 3.7)" .IR exclude_callchain_user " (since Linux 3.7)"
.\" commit d077526485d5c9b12fe85d0b2b3b7041e6bc5f91 .\" commit d077526485d5c9b12fe85d0b2b3b7041e6bc5f91
Do not include user callchains. Do not include user callchains.
.TP .TP
.IR "mmap2" " (since Linux 3.16)" .IR mmap2 " (since Linux 3.16)"
.\" commit 13d7a2410fa637f450a29ecb515ac318ee40c741 .\" commit 13d7a2410fa637f450a29ecb515ac318ee40c741
.\" This is tricky; was committed during 3.12 development .\" This is tricky; was committed during 3.12 development
.\" but right before release was disabled. .\" but right before release was disabled.
@ -1152,7 +1162,7 @@ The
.I mmap .I mmap
flag must also be set for this to work. flag must also be set for this to work.
.TP .TP
.IR "comm_exec" " (since Linux 3.16)" .IR comm_exec " (since Linux 3.16)"
.\" commit 82b897782d10fcc4930c9d4a15b175348fdd2871 .\" commit 82b897782d10fcc4930c9d4a15b175348fdd2871
This is purely a feature-detection flag, it does not change This is purely a feature-detection flag, it does not change
kernel behavior. kernel behavior.
@ -1168,7 +1178,7 @@ reported was caused by a call to
This allows tools to distinguish between the various This allows tools to distinguish between the various
types of process renaming. types of process renaming.
.TP .TP
.IR "use_clockid" " (since Linux 4.1)" .IR use_clockid " (since Linux 4.1)"
.\" commit 34f439278cef7b1177f8ce24f9fc81dfc6221d3b .\" commit 34f439278cef7b1177f8ce24f9fc81dfc6221d3b
This allows selecting which internal Linux clock to use This allows selecting which internal Linux clock to use
when generating timestamps via the when generating timestamps via the
@ -1177,7 +1187,7 @@ field.
This can make it easier to correlate perf sample times with This can make it easier to correlate perf sample times with
timestamps generated by other tools. timestamps generated by other tools.
.TP .TP
.IR "context_switch" " (since Linux 4.3)" .IR context_switch " (since Linux 4.3)"
.\" commit 45ac1403f564f411c6a383a2448688ba8dd705a4 .\" commit 45ac1403f564f411c6a383a2448688ba8dd705a4
This enables the generation of This enables the generation of
.B PERF_RECORD_SWITCH .B PERF_RECORD_SWITCH
@ -1192,7 +1202,7 @@ information even with strict
.I perf_event_paranoid .I perf_event_paranoid
settings. settings.
.TP .TP
.IR "wakeup_events" ", " "wakeup_watermark" .IR wakeup_events ", " wakeup_watermark
This union sets how many samples This union sets how many samples
.RI ( wakeup_events ) .RI ( wakeup_events )
or bytes or bytes
@ -1218,25 +1228,25 @@ Prior to Linux 3.0, setting
to 0 resulted in no overflow notifications; to 0 resulted in no overflow notifications;
more recent kernels treat 0 the same as 1. more recent kernels treat 0 the same as 1.
.TP .TP
.IR "bp_type" " (since Linux 2.6.33)" .IR bp_type " (since Linux 2.6.33)"
.\" commit 24f1e32c60c45c89a997c73395b69c8af6f0a84e .\" commit 24f1e32c60c45c89a997c73395b69c8af6f0a84e
This chooses the breakpoint type. This chooses the breakpoint type.
It is one of: It is one of:
.RS .RS
.TP .TP
.BR HW_BREAKPOINT_EMPTY .B HW_BREAKPOINT_EMPTY
No breakpoint. No breakpoint.
.TP .TP
.BR HW_BREAKPOINT_R .B HW_BREAKPOINT_R
Count when we read the memory location. Count when we read the memory location.
.TP .TP
.BR HW_BREAKPOINT_W .B HW_BREAKPOINT_W
Count when we write the memory location. Count when we write the memory location.
.TP .TP
.BR HW_BREAKPOINT_RW .B HW_BREAKPOINT_RW
Count when we read or write the memory location. Count when we read or write the memory location.
.TP .TP
.BR HW_BREAKPOINT_X .B HW_BREAKPOINT_X
Count when we execute code at the memory location. Count when we execute code at the memory location.
.PP .PP
The values can be combined via a bitwise or, but the The values can be combined via a bitwise or, but the
@ -1249,14 +1259,14 @@ with
is not allowed. is not allowed.
.RE .RE
.TP .TP
.IR "bp_addr" " (since Linux 2.6.33)" .IR bp_addr " (since Linux 2.6.33)"
.\" commit 24f1e32c60c45c89a997c73395b69c8af6f0a84e .\" commit 24f1e32c60c45c89a997c73395b69c8af6f0a84e
This is the address of the breakpoint. This is the address of the breakpoint.
For execution breakpoints, this is the memory address of the instruction For execution breakpoints, this is the memory address of the instruction
of interest; for read and write breakpoints, it is the memory address of interest; for read and write breakpoints, it is the memory address
of the memory location of interest. of the memory location of interest.
.TP .TP
.IR "config1" " (since Linux 2.6.39)" .IR config1 " (since Linux 2.6.39)"
.\" commit a7e3ed1e470116c9d12c2f778431a481a6be8ab6 .\" commit a7e3ed1e470116c9d12c2f778431a481a6be8ab6
.I config1 .I config1
is used for setting events that need an extra register or otherwise is used for setting events that need an extra register or otherwise
@ -1264,7 +1274,7 @@ do not fit in the regular config field.
Raw OFFCORE_EVENTS on Nehalem/Westmere/SandyBridge use this field Raw OFFCORE_EVENTS on Nehalem/Westmere/SandyBridge use this field
on Linux 3.3 and later kernels. on Linux 3.3 and later kernels.
.TP .TP
.IR "bp_len" " (since Linux 2.6.33)" .IR bp_len " (since Linux 2.6.33)"
.\" commit 24f1e32c60c45c89a997c73395b69c8af6f0a84e .\" commit 24f1e32c60c45c89a997c73395b69c8af6f0a84e
.I bp_len .I bp_len
is the length of the breakpoint being measured if is the length of the breakpoint being measured if
@ -1280,14 +1290,14 @@ and
For an execution breakpoint, set this to For an execution breakpoint, set this to
.IR sizeof(long) . .IR sizeof(long) .
.TP .TP
.IR "config2" " (since Linux 2.6.39)" .IR config2 " (since Linux 2.6.39)"
.\" commit a7e3ed1e470116c9d12c2f778431a481a6be8ab6 .\" commit a7e3ed1e470116c9d12c2f778431a481a6be8ab6
.I config2 .I config2
is a further extension of the is a further extension of the
.I config1 .I config1
field. field.
.TP .TP
.IR "branch_sample_type" " (since Linux 3.4)" .IR branch_sample_type " (since Linux 3.4)"
.\" commit bce38cd53e5ddba9cb6d708c4ef3d04a4016ec7e .\" commit bce38cd53e5ddba9cb6d708c4ef3d04a4016ec7e
If If
.B PERF_SAMPLE_BRANCH_STACK .B PERF_SAMPLE_BRANCH_STACK
@ -1358,20 +1368,20 @@ This requires hardware support, currently only found
on Intel x86 Haswell or newer. on Intel x86 Haswell or newer.
.RE .RE
.TP .TP
.IR "sample_regs_user" " (since Linux 3.7)" .IR sample_regs_user " (since Linux 3.7)"
.\" commit 4018994f3d8785275ef0e7391b75c3462c029e56 .\" commit 4018994f3d8785275ef0e7391b75c3462c029e56
This bit mask defines the set of user CPU registers to dump on samples. This bit mask defines the set of user CPU registers to dump on samples.
The layout of the register mask is architecture-specific and The layout of the register mask is architecture-specific and
is described in the kernel header file is described in the kernel header file
.IR arch/ARCH/include/uapi/asm/perf_regs.h . .IR arch/ARCH/include/uapi/asm/perf_regs.h .
.TP .TP
.IR "sample_stack_user" " (since Linux 3.7)" .IR sample_stack_user " (since Linux 3.7)"
.\" commit c5ebcedb566ef17bda7b02686e0d658a7bb42ee7 .\" commit c5ebcedb566ef17bda7b02686e0d658a7bb42ee7
This defines the size of the user stack to dump if This defines the size of the user stack to dump if
.B PERF_SAMPLE_STACK_USER .B PERF_SAMPLE_STACK_USER
is specified. is specified.
.TP .TP
.IR "clockid" " (since Linux 4.1)" .IR clockid " (since Linux 4.1)"
.\" commit 34f439278cef7b1177f8ce24f9fc81dfc6221d3b .\" commit 34f439278cef7b1177f8ce24f9fc81dfc6221d3b
If If
.I use_clockid .I use_clockid
@ -1388,13 +1398,13 @@ and
.B CLOCK_TAI .B CLOCK_TAI
currently supported. currently supported.
.TP .TP
.IR "aux_watermark" " (since Linux 4.1)" .IR aux_watermark " (since Linux 4.1)"
.\" commit 1a5941312414c71dece6717da9a0fa1303127afa .\" commit 1a5941312414c71dece6717da9a0fa1303127afa
This specifies how much data is required to trigger a This specifies how much data is required to trigger a
.B PERF_RECORD_AUX .B PERF_RECORD_AUX
sample. sample.
.TP .TP
.IR "sample_max_stack" " (since Linux 4.8)" .IR sample_max_stack " (since Linux 4.8)"
.\" commit 97c79a38cd454602645f0470ffb444b3b75ce574 .\" commit 97c79a38cd454602645f0470ffb444b3b75ce574
When When
.I sample_type .I sample_type
@ -1682,34 +1692,39 @@ delta since
.I time_enabled .I time_enabled
(in nanoseconds) using rdtsc or similar. (in nanoseconds) using rdtsc or similar.
.IP .IP
.nf .in +4n
u64 quot, rem; .EX
u64 delta; u64 quot, rem;
quot = (cyc >> time_shift); u64 delta;
rem = cyc & (((u64)1 << time_shift) \- 1);
delta = time_offset + quot * time_mult + quot = cyc >> time_shift;
((rem * time_mult) >> time_shift); rem = cyc & (((u64)1 << time_shift) \- 1);
.fi delta = time_offset + quot * time_mult +
((rem * time_mult) >> time_shift);
.EE
.in
.IP .IP
Where Where
.IR time_offset , .IR time_offset ,
.IR time_mult , .IR time_mult ,
.IR time_shift , .IR time_shift ,
and and
.IR cyc .I cyc
are read in the are read in the
seqcount loop described above. seqcount loop described above.
This delta can then be added to This delta can then be added to
enabled and possible running (if idx), improving the scaling: enabled and possible running (if idx), improving the scaling:
.IP .IP
.nf .in +4n
enabled += delta; .EX
if (idx) enabled += delta;
running += delta; if (idx)
quot = count / running; running += delta;
rem = count % running; quot = count / running;
count = quot * enabled + (rem * enabled) / running; rem = count % running;
.fi count = quot * enabled + (rem * enabled) / running;
.EE
.in
.TP .TP
.IR time_zero " (since Linux 3.12)" .IR time_zero " (since Linux 3.12)"
.\" commit fa7315871046b9a4c48627905691dbde57e51033 .\" commit fa7315871046b9a4c48627905691dbde57e51033
@ -1718,23 +1733,31 @@ If
.I cap_usr_time_zero .I cap_usr_time_zero
is set, then the hardware clock (the TSC timestamp counter on x86) is set, then the hardware clock (the TSC timestamp counter on x86)
can be calculated from the can be calculated from the
.IR time_zero ", " time_mult ", and " time_shift " values:" .IR time_zero ,
.IR time_mult ,
and
.I time_shift
values:
.IP .IP
.nf .in +4n
time = timestamp - time_zero; .EX
quot = time / time_mult; time = timestamp - time_zero;
rem = time % time_mult; quot = time / time_mult;
cyc = (quot << time_shift) + (rem << time_shift) / time_mult; rem = time % time_mult;
.fi cyc = (quot << time_shift) + (rem << time_shift) / time_mult;
.EE
.in
.IP .IP
And vice versa: And vice versa:
.IP .IP
.nf .in +4n
quot = cyc >> time_shift; .EX
rem = cyc & (((u64)1 << time_shift) - 1); quot = cyc >> time_shift;
timestamp = time_zero + quot * time_mult + rem = cyc & (((u64)1 << time_shift) - 1);
((rem * time_mult) >> time_shift); timestamp = time_zero + quot * time_mult +
.fi ((rem * time_mult) >> time_shift);
.EE
.in
.TP .TP
.I data_head .I data_head
This points to the head of the data section. This points to the head of the data section.
@ -1785,7 +1808,7 @@ The desired offset and size must be page aligned, and the size
must be a power of two. must be a power of two.
These values are then passed to mmap in order to map the AUX buffer. 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 Pages in the AUX buffer are included as part of the
.BR RLIMIT_MEMLOCK .B RLIMIT_MEMLOCK
resource limit (see resource limit (see
.BR setrlimit (2)), .BR setrlimit (2)),
and also as part of the and also as part of the
@ -1802,10 +1825,14 @@ new data began, and it is the consumer's job to disable
measurement while reading to avoid possible data races. measurement while reading to avoid possible data races.
.IP .IP
The The
.IR aux_head " and " aux_tail .I aux_head
and
.I aux_tail
ring buffer pointers have the same behavior and ordering ring buffer pointers have the same behavior and ordering
rules as the previous described rules as the previous described
.IR data_head " and " data_tail . .I data_head
and
.IR data_tail .
.PP .PP
The following 2^n ring-buffer pages have the layout described below. The following 2^n ring-buffer pages have the layout described below.
.PP .PP
@ -1897,9 +1924,9 @@ system call.
.BR PERF_RECORD_MISC_SWITCH_OUT " (since Linux 4.3)" .BR PERF_RECORD_MISC_SWITCH_OUT " (since Linux 4.3)"
.\" commit 45ac1403f564f411c6a383a2448688ba8dd705a4 .\" commit 45ac1403f564f411c6a383a2448688ba8dd705a4
When a When a
.BR PERF_RECORD_SWITCH .B PERF_RECORD_SWITCH
or or
.BR PERF_RECORD_SWITCH_CPU_WIDE .B PERF_RECORD_SWITCH_CPU_WIDE
record is generated, this bit indicates that the record is generated, this bit indicates that the
context switch is away from the current process context switch is away from the current process
(instead of into the current process). (instead of into the current process).
@ -2267,7 +2294,7 @@ Support for
.IR mispred , .IR mispred ,
.IR predicted , .IR predicted ,
and and
.IR cycles .I cycles
is optional; if not supported, those is optional; if not supported, those
values will be 0. values will be 0.
.PP .PP
@ -2284,7 +2311,9 @@ is enabled, then the user CPU registers are recorded.
The The
.I abi .I abi
field is one of field is one of
.BR PERF_SAMPLE_REGS_ABI_NONE ", " PERF_SAMPLE_REGS_ABI_32 " or" .BR PERF_SAMPLE_REGS_ABI_NONE ,
.BR PERF_SAMPLE_REGS_ABI_32 ,
or
.BR PERF_SAMPLE_REGS_ABI_64 . .BR PERF_SAMPLE_REGS_ABI_64 .
.IP .IP
The The