tjh.dev
Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
perf mem: Update documentation for new options
Add a common options section and move some items to the section. Also
add description of new options to report options.
Suggested-by: Ian Rogers <irogers@google.com>
Reviewed-by: Ian Rogers <irogers@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Athira Rajeev <atrajeev@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: https://lore.kernel.org/lkml/20240802180913.1023886-1-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
authored by
Namhyung Kim
and committed by
Arnaldo Carvalho de Melo
13159a13
7320ad97
+63
-31
+63
-31
tools/perf/Documentation/perf-mem.txt
+63
-31
tools/perf/Documentation/perf-mem.txt
···
28
28
Due to the statistical nature of SPE sampling, not every memory operation will
29
29
be sampled.
30
30
31
-
OPTIONS
32
-
-------
33
-
<command>...::
34
-
Any command you can specify in a shell.
35
-
36
-
-i::
37
-
--input=<file>::
38
-
Input file name.
39
-
31
+
COMMON OPTIONS
32
+
--------------
40
33
-f::
41
34
--force::
42
35
Don't do ownership validation
···
38
45
--type=<type>::
39
46
Select the memory operation type: load or store (default: load,store)
40
47
41
-
-D::
42
-
--dump-raw-samples::
43
-
Dump the raw decoded samples on the screen in a format that is easy to parse with
44
-
one sample per line.
45
-
46
-
-x::
47
-
--field-separator=<separator>::
48
-
Specify the field separator used when dump raw samples (-D option). By default,
49
-
The separator is the space character.
50
-
51
-
-C::
52
-
--cpu=<cpu>::
53
-
Monitor only on the list of CPUs provided. Multiple CPUs can be provided as a
54
-
comma-separated list with no space: 0,1. Ranges of CPUs are specified with -: 0-2. Default
55
-
is to monitor all CPUS.
56
-
-U::
57
-
--hide-unresolved::
58
-
Only display entries resolved to a symbol.
48
+
-v::
49
+
--verbose::
50
+
Be more verbose (show counter open errors, etc)
59
51
60
52
-p::
61
53
--phys-data::
···
51
73
52
74
RECORD OPTIONS
53
75
--------------
76
+
<command>...::
77
+
Any command you can specify in a shell.
78
+
54
79
-e::
55
80
--event <event>::
56
81
Event selector. Use 'perf mem record -e list' to list available events.
···
66
85
--all-user::
67
86
Configure all used events to run in user space.
68
87
69
-
-v::
70
-
--verbose::
71
-
Be more verbose (show counter open errors, etc)
72
-
73
88
--ldlat <n>::
74
89
Specify desired latency for loads event. Supported on Intel and Arm64
75
90
processors only. Ignored on other archs.
91
+
92
+
REPORT OPTIONS
93
+
--------------
94
+
-i::
95
+
--input=<file>::
96
+
Input file name.
97
+
98
+
-C::
99
+
--cpu=<cpu>::
100
+
Monitor only on the list of CPUs provided. Multiple CPUs can be provided as a
101
+
comma-separated list with no space: 0,1. Ranges of CPUs are specified with -
102
+
like 0-2. Default is to monitor all CPUS.
103
+
104
+
-D::
105
+
--dump-raw-samples::
106
+
Dump the raw decoded samples on the screen in a format that is easy to parse with
107
+
one sample per line.
108
+
109
+
-s::
110
+
--sort=<key>::
111
+
Group result by given key(s) - multiple keys can be specified
112
+
in CSV format. The keys are specific to memory samples are:
113
+
symbol_daddr, symbol_iaddr, dso_daddr, locked, tlb, mem, snoop,
114
+
dcacheline, phys_daddr, data_page_size, blocked.
115
+
116
+
- symbol_daddr: name of data symbol being executed on at the time of sample
117
+
- symbol_iaddr: name of code symbol being executed on at the time of sample
118
+
- dso_daddr: name of library or module containing the data being executed
119
+
on at the time of the sample
120
+
- locked: whether the bus was locked at the time of the sample
121
+
- tlb: type of tlb access for the data at the time of the sample
122
+
- mem: type of memory access for the data at the time of the sample
123
+
- snoop: type of snoop (if any) for the data at the time of the sample
124
+
- dcacheline: the cacheline the data address is on at the time of the sample
125
+
- phys_daddr: physical address of data being executed on at the time of sample
126
+
- data_page_size: the data page size of data being executed on at the time of sample
127
+
- blocked: reason of blocked load access for the data at the time of the sample
128
+
129
+
And the default sort keys are changed to local_weight, mem, sym, dso,
130
+
symbol_daddr, dso_daddr, snoop, tlb, locked, blocked, local_ins_lat.
131
+
132
+
-T::
133
+
--type-profile::
134
+
Show data-type profile result instead of code symbols. This requires
135
+
the debug information and it will change the default sort keys to:
136
+
mem, snoop, tlb, type.
137
+
138
+
-U::
139
+
--hide-unresolved::
140
+
Only display entries resolved to a symbol.
141
+
142
+
-x::
143
+
--field-separator=<separator>::
144
+
Specify the field separator used when dump raw samples (-D option). By default,
145
+
The separator is the space character.
76
146
77
147
In addition, for report all perf report options are valid, and for record
78
148
all perf record options.