| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
The num_print_interval and config->interval_clear should be checked
together like other places like later in the function. Otherwise,
the --interval-clear option could print the headers for the CSV or
JSON output unnecessarily.
Signed-off-by: Namhyung Kim <[email protected]>
Acked-by: Ian Rogers <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
It missed to print a matching header line for intervals.
Before:
# perf stat -a -e cycles,instructions --metric-only -j -I 500
{"unit" : "insn per cycle"}
{"interval" : 0.500544283}{"metric-value" : "1.96"}
^C
After:
# perf stat -a -e cycles,instructions --metric-only -j -I 500
{"unit" : "sec"}{"unit" : "insn per cycle"}
{"interval" : 0.500515681}{"metric-value" : "2.31"}
^C
Signed-off-by: Namhyung Kim <[email protected]>
Acked-by: Ian Rogers <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Currently --metric-only with --json indents header lines. This is not
needed for JSON.
$ perf stat -aA --metric-only -j true
{"unit" : "GHz"}{"unit" : "insn per cycle"}{"unit" : "branch-misses of all branches"}
{"cpu" : "0", {"metric-value" : "0.101"}{"metric-value" : "0.86"}{"metric-value" : "1.91"}
{"cpu" : "1", {"metric-value" : "0.102"}{"metric-value" : "0.87"}{"metric-value" : "2.02"}
{"cpu" : "2", {"metric-value" : "0.085"}{"metric-value" : "1.02"}{"metric-value" : "1.69"}
...
Note that the other lines are broken JSON, but it will be handled later.
Signed-off-by: Namhyung Kim <[email protected]>
Acked-by: Ian Rogers <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Currently it prints all metric headers for JSON output. But actually it
skips some metrics with valid_only_metric(). So the output looks like:
$ perf stat --metric-only --json true
{"unit" : "CPUs utilized", "unit" : "/sec", "unit" : "/sec", "unit" : "/sec", "unit" : "GHz", "unit" : "insn per cycle", "unit" : "/sec", "unit" : "branch-misses of all branches"}
{"metric-value" : "3.861"}{"metric-value" : "0.79"}{"metric-value" : "3.04"}
As you can see there are 8 units in the header but only 3 metric-values
are there. It should skip the unused headers as well. Also each unit
should be printed as a separate object like metric values.
With this patch:
$ perf stat --metric-only --json true
{"unit" : "GHz"}{"unit" : "insn per cycle"}{"unit" : "branch-misses of all branches"}
{"metric-value" : "4.166"}{"metric-value" : "0.73"}{"metric-value" : "2.96"}
Fixes: df936cadfb58ba93 ("perf stat: Add JSON output option")
Signed-off-by: Namhyung Kim <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Claire Jensen <[email protected]>
Cc: Ian Rogers <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
The struct perf_stat_output_ctx is set in a loop with the same values.
Move the code out of the loop and keep the loop minimal.
Signed-off-by: Namhyung Kim <[email protected]>
Acked-by: Ian Rogers <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
The --interval-clear option makes perf stat to clear the terminal at
each interval. But it doesn't need to clear the screen when it saves
to a file.
Signed-off-by: Namhyung Kim <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Ian Rogers <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
When perf stat is called with very detailed events, the output doesn't
align well like below:
$ sudo perf stat -a -ddd sleep 1
Performance counter stats for 'system wide':
8,020.23 msec cpu-clock # 7.997 CPUs utilized
3,970 context-switches # 494.998 /sec
169 cpu-migrations # 21.072 /sec
586 page-faults # 73.065 /sec
649,568,060 cycles # 0.081 GHz (30.42%)
304,044,345 instructions # 0.47 insn per cycle (38.40%)
60,313,022 branches # 7.520 M/sec (38.89%)
2,766,919 branch-misses # 4.59% of all branches (39.26%)
74,422,951 L1-dcache-loads # 9.279 M/sec (39.39%)
8,025,568 L1-dcache-load-misses # 10.78% of all L1-dcache accesses (39.22%)
3,314,995 LLC-loads # 413.329 K/sec (30.83%)
1,225,619 LLC-load-misses # 36.97% of all LL-cache accesses (30.45%)
<not supported> L1-icache-loads
20,420,493 L1-icache-load-misses # 0.00% of all L1-icache accesses (30.29%)
58,017,947 dTLB-loads # 7.234 M/sec (30.37%)
704,677 dTLB-load-misses # 1.21% of all dTLB cache accesses (30.27%)
234,225 iTLB-loads # 29.204 K/sec (30.29%)
417,166 iTLB-load-misses # 178.10% of all iTLB cache accesses (30.32%)
<not supported> L1-dcache-prefetches
<not supported> L1-dcache-prefetch-misses
1.002947355 seconds time elapsed
Increase the METRIC_LEN by 3 so that it can align properly.
Signed-off-by: Namhyung Kim <[email protected]>
Acked-by: Ian Rogers <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
A fix for the LLVM compilation error while building bpftool.
Replaces the expression:
_Static_assert((p) == NULL || ...)
by expression:
_Static_assert((__builtin_constant_p((p)) ? (p) == NULL : 0) || ...)
When "p" is not a constant the former is not considered to be a
constant expression by LLVM 14.
The error was introduced in the following patch-set: [1].
The error was reported here: [2].
[1] https://lore.kernel.org/bpf/[email protected]/
[2] https://lore.kernel.org/all/[email protected]/
Reported-by: kernel test robot <[email protected]>
Fixes: c302378bc157 ("libbpf: Hashmap interface update to allow both long and void* keys/values")
Signed-off-by: Eduard Zingerman <[email protected]>
Signed-off-by: Andrii Nakryiko <[email protected]>
Acked-by: Stanislav Fomichev <[email protected]>
Link: https://lore.kernel.org/bpf/[email protected]
|
|
Since variable npmus is unsigned int, comparing with 0 is unnecessary.
Signed-off-by: Kang Minchul <[email protected]>
Acked-by: Namhyung Kim <[email protected]>
Cc: Alexander Shishkin <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Mark Rutland <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
When converting recorded data into JSON format, perf data omits probe
variables. Add them to the output in the format "field name": "field value"
using tep_print_field:
$ perf data convert --to-json output.json
// output.json
{
"linux-perf-json-version": 1,
"headers": { ... },
"samples": [
{
"timestamp": 29182079082999,
"pid": 309194,
[...]
"__probe_ip": "0x93ee35",
"query_string_string": "select 2;",
"nxids": "0"
}
]
}
Signed-off-by: Dmitrii Dolgov <[email protected]>
Acked-by: Ian Rogers <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jiri Olsa <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
An update for libbpf's hashmap interface from void* -> void* to a
polymorphic one, allowing both long and void* keys and values.
This simplifies many use cases in libbpf as hashmaps there are mostly
integer to integer.
Perf copies hashmap implementation from libbpf and has to be
updated as well.
Changes to libbpf, selftests/bpf and perf are packed as a single
commit to avoid compilation issues with any future bisect.
Polymorphic interface is acheived by hiding hashmap interface
functions behind auxiliary macros that take care of necessary
type casts, for example:
#define hashmap_cast_ptr(p) \
({ \
_Static_assert((p) == NULL || sizeof(*(p)) == sizeof(long),\
#p " pointee should be a long-sized integer or a pointer"); \
(long *)(p); \
})
bool hashmap_find(const struct hashmap *map, long key, long *value);
#define hashmap__find(map, key, value) \
hashmap_find((map), (long)(key), hashmap_cast_ptr(value))
- hashmap__find macro casts key and value parameters to long
and long* respectively
- hashmap_cast_ptr ensures that value pointer points to a memory
of appropriate size.
This hack was suggested by Andrii Nakryiko in [1].
This is a follow up for [2].
[1] https://lore.kernel.org/bpf/CAEf4BzZ8KFneEJxFAaNCCFPGqp20hSpS2aCj76uRk3-qZUH5xg@mail.gmail.com/
[2] https://lore.kernel.org/bpf/[email protected]/T/#m65b28f1d6d969fcd318b556db6a3ad499a42607d
Signed-off-by: Eduard Zingerman <[email protected]>
Signed-off-by: Andrii Nakryiko <[email protected]>
Link: https://lore.kernel.org/bpf/[email protected]
|
|
DWARF version 5 standard Sec 2.14 says that
Any debugging information entry representing the declaration of an object,
module, subprogram or type may have DW_AT_decl_file, DW_AT_decl_line and
DW_AT_decl_column attributes, each of whose value is an unsigned integer
constant.
So it should be an unsigned integer data. Also, even though the standard
doesn't clearly say the DW_AT_call_file is signed or unsigned, the
elfutils (eu-readelf) interprets it as unsigned integer data and it is
natural to handle it as unsigned integer data as same as DW_AT_decl_file.
This changes the DW_AT_call_file as unsigned integer data too.
Fixes: 3f4460a28fb2f73d ("perf probe: Filter out redundant inline-instances")
Signed-off-by: Masami Hiramatsu <[email protected]>
Acked-by: Namhyung Kim <[email protected]>
Cc: Alexander Shishkin <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Mark Rutland <[email protected]>
Cc: Masami Hiramatsu <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: [email protected]
Cc: Steven Rostedt (VMware) <[email protected]>
Link: https://lore.kernel.org/r/166761727445.480106.3738447577082071942.stgit@devnote3
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Commit f4a2aade6809c657 ("perf tests powerpc: Fix branch stack sampling
test to include sanity check for branch filter") added a skip if certain
branch options aren't available.
But the change added both -b (--branch-any) and --branch-filter options
at the same time, which will always result in a failure on any platform
because the arguments can't be used together.
Fix this by removing -b (--branch-any) and leaving --branch-filter which
already specifies 'any'. Also add warning messages to the test and perf
tool.
Output on x86 before this fix:
$ sudo ./perf test branch
108: Check branch stack sampling : Skip
After:
$ sudo ./perf test branch
108: Check branch stack sampling : Ok
Fixes: f4a2aade6809c657 ("perf tests powerpc: Fix branch stack sampling test to include sanity check for branch filter")
Signed-off-by: James Clark <[email protected]>
Tested-by: Athira Jajeev <[email protected]>
Cc: Alexander Shishkin <[email protected]>
Cc: [email protected]
Cc: Ingo Molnar <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kajol Jain <[email protected]>
Cc: Mark Rutland <[email protected]>
Cc: Namhyung Kim <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
'perf stat' with CSV output option prints an extra empty string as first
field in metrics output line. Sample output below:
# ./perf stat -x, --per-socket -a -C 1 ls
S0,1,1.78,msec,cpu-clock,1785146,100.00,0.973,CPUs utilized
S0,1,26,,context-switches,1781750,100.00,0.015,M/sec
S0,1,1,,cpu-migrations,1780526,100.00,0.561,K/sec
S0,1,1,,page-faults,1779060,100.00,0.561,K/sec
S0,1,875807,,cycles,1769826,100.00,0.491,GHz
S0,1,85281,,stalled-cycles-frontend,1767512,100.00,9.74,frontend cycles idle
S0,1,576839,,stalled-cycles-backend,1766260,100.00,65.86,backend cycles idle
S0,1,288430,,instructions,1762246,100.00,0.33,insn per cycle
====> ,S0,1,,,,,,,2.00,stalled cycles per insn
The above command line uses field separator as "," via "-x," option and
per-socket option displays socket value as first field. But here the
last line for "stalled cycles per insn" has "," in the beginning.
Sample output using interval mode:
# ./perf stat -I 1000 -x, --per-socket -a -C 1 ls
0.001813453,S0,1,1.87,msec,cpu-clock,1872052,100.00,0.002,CPUs utilized
0.001813453,S0,1,2,,context-switches,1868028,100.00,1.070,K/sec
------
0.001813453,S0,1,85379,,instructions,1856754,100.00,0.32,insn per cycle
====> 0.001813453,,S0,1,,,,,,,1.34,stalled cycles per insn
Above result also has an extra CSV separator after
the timestamp. Patch addresses extra field separator
in the beginning of the metric output line.
The counter stats are displayed by function
"perf_stat__print_shadow_stats" in code
"util/stat-shadow.c". While printing the stats info
for "stalled cycles per insn", function "new_line_csv"
is used as new_line callback.
The new_line_csv function has check for "os->prefix"
and if prefix is not null, it will be printed along
with cvs separator.
Snippet from "new_line_csv":
if (os->prefix)
fprintf(os->fh, "%s%s", os->prefix, config->csv_sep);
Here os->prefix gets printed followed by ","
which is the cvs separator. The os->prefix is
used in interval mode option ( -I ), to print
time stamp on every new line. But prefix is
already set to contain CSV separator when used
in interval mode for CSV option.
Reference: Function "static void print_interval"
Snippet:
sprintf(prefix, "%6lu.%09lu%s", ts->tv_sec, ts->tv_nsec, config->csv_sep);
Also if prefix is not assigned (if not used with
-I option), it gets set to empty string.
Reference: function printout() in util/stat-display.c
Snippet:
.prefix = prefix ? prefix : "",
Since prefix already set to contain cvs_sep in interval
option, patch removes printing config->csv_sep in
new_line_csv function to avoid printing extra field.
After the patch:
# ./perf stat -x, --per-socket -a -C 1 ls
S0,1,2.04,msec,cpu-clock,2045202,100.00,1.013,CPUs utilized
S0,1,2,,context-switches,2041444,100.00,979.289,/sec
S0,1,0,,cpu-migrations,2040820,100.00,0.000,/sec
S0,1,2,,page-faults,2040288,100.00,979.289,/sec
S0,1,254589,,cycles,2036066,100.00,0.125,GHz
S0,1,82481,,stalled-cycles-frontend,2032420,100.00,32.40,frontend cycles idle
S0,1,113170,,stalled-cycles-backend,2031722,100.00,44.45,backend cycles idle
S0,1,88766,,instructions,2030942,100.00,0.35,insn per cycle
S0,1,,,,,,,1.27,stalled cycles per insn
Fixes: 92a61f6412d3a09d ("perf stat: Implement CSV metrics output")
Reported-by: Disha Goel <[email protected]>
Reviewed-By: Kajol Jain <[email protected]>
Signed-off-by: Athira Jajeev <[email protected]>
Tested-by: Disha Goel <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Ian Rogers <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: [email protected]
Cc: Madhavan Srinivasan <[email protected]>
Cc: Michael Ellerman <[email protected]>
Cc: Nageswara R Sastry <[email protected]>
Cc: Namhyung Kim <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
The following command will get segfault due to missing aggr_header_csv
for AGGR_NODE:
$ sudo perf stat -a --per-node -x, --metric-only true
Committer testing:
Before this patch:
# perf stat -a --per-node -x, --metric-only true
Segmentation fault (core dumped)
#
After:
# gdb perf
-bash: gdb: command not found
# perf stat -a --per-node -x, --metric-only true
node,Ghz,frontend cycles idle,backend cycles idle,insn per cycle,branch-misses of all branches,
N0,32,0.335,2.10,0.65,0.69,0.03,1.92,
#
Fixes: 86895b480a2f10c7 ("perf stat: Add --per-node agregation support")
Signed-off-by: Namhyung Kim <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Ian Rogers <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: http://lore.kernel.org/lkml/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
As this is where we expect to find bpf/bpf_helpers.h, etc.
This needs more work to make it follow LIBBPF_DYNAMIC=1 usage, i.e. when
not using the system libbpf it should use the headers in the in-kernel
sources libbpf in tools/lib/bpf.
We need to do that anyway to avoid this mixup system libbpf and
in-kernel files, so we'll get this sorted out that way.
And this also may become moot as we move to using BPF skels for this
feature.
Cc: Adrian Hunter <[email protected]>
Cc: Ian Rogers <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Namhyung Kim <[email protected]>
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Don't use deprecated and now broken map style. Avoid use of
tools/perf/include/bpf/bpf.h and use the more regular BPF headers.
Committer notes:
Add /usr/include to the include path so that bpf/bpf_helpers.h can be
found, remove sys/socket.h, adding the sockaddr_storage definition, also
remove stdbool.h, both were preventing building the
augmented_raw_syscalls.c file with clang, revisit later.
Testing it:
Asking for syscalls that have string arguments:
# perf trace -e ~acme/git/perf/tools/perf/examples/bpf/augmented_raw_syscalls.c,string --max-events 10
0.000 thermald/1144 openat(dfd: CWD, filename: "/sys/class/powercap/intel-rapl/intel-rapl:0/intel-rapl:0:2/energy_uj", flags: RDONLY) = 13
0.158 thermald/1144 openat(dfd: CWD, filename: "/sys/class/powercap/intel-rapl/intel-rapl:0/energy_uj", flags: RDONLY) = 13
0.215 thermald/1144 openat(dfd: CWD, filename: "/sys/class/thermal/thermal_zone3/temp", flags: RDONLY) = 13
16.448 cgroupify/36478 openat(dfd: 4, filename: ".", flags: RDONLY|CLOEXEC|DIRECTORY|NONBLOCK) = 5
16.468 cgroupify/36478 newfstatat(dfd: 5, filename: "", statbuf: 0x7fffca5b4130, flag: 4096) = 0
16.473 systemd-oomd/972 openat(dfd: CWD, filename: "/proc/meminfo", flags: RDONLY|CLOEXEC) = 12
16.499 systemd-oomd/972 newfstatat(dfd: 12, filename: "", statbuf: 0x7ffd2bc73cc0, flag: 4096) = 0
16.516 abrt-dump-jour/1370 openat(dfd: CWD, filename: "/var/log/journal/d6a97235307247e09f13f326fb607e3c/system.journal", flags: RDONLY|CLOEXEC|NONBLOCK) = 21
16.538 abrt-dump-jour/1370 newfstatat(dfd: 21, filename: "", statbuf: 0x7ffc651b8980, flag: 4096) = 0
16.540 abrt-dump-jour/1371 openat(dfd: CWD, filename: "/var/log/journal/d6a97235307247e09f13f326fb607e3c/system.journal", flags: RDONLY|CLOEXEC|NONBLOCK) = 21
#
Networking syscalls:
# perf trace -e ~acme/git/perf/tools/perf/examples/bpf/augmented_raw_syscalls.c,sendto*,connect* --max-events 10
0.000 isc-net-0005/1206 connect(fd: 512, uservaddr: { .family: INET, port: 53, addr: 23.211.132.65 }, addrlen: 16) = 0
0.070 isc-net-0002/1203 connect(fd: 515, uservaddr: { .family: INET6, port: 53, addr: 2600:1401:2::43 }, addrlen: 28) = -1 ENETUNREACH (Network is unreachable)
0.031 isc-net-0006/1207 connect(fd: 513, uservaddr: { .family: INET6, port: 53, addr: 2600:1401:2::43 }, addrlen: 28) = -1 ENETUNREACH (Network is unreachable)
0.079 isc-net-0006/1207 sendto(fd: 3, buff: 0x7f73a40611b0, len: 106, flags: NOSIGNAL, addr: { .family: UNSPEC }, addr_len: NULL) = 106
0.180 isc-net-0006/1207 connect(fd: 519, uservaddr: { .family: INET6, port: 53, addr: 2600:1401:1::43 }, addrlen: 28) = -1 ENETUNREACH (Network is unreachable)
0.211 isc-net-0006/1207 sendto(fd: 3, buff: 0x7f73a4061230, len: 106, flags: NOSIGNAL, addr: { .family: UNSPEC }, addr_len: NULL) = 106
0.298 isc-net-0006/1207 connect(fd: 515, uservaddr: { .family: INET, port: 53, addr: 96.7.49.67 }, addrlen: 16) = 0
0.109 isc-net-0004/1205 connect(fd: 518, uservaddr: { .family: INET6, port: 53, addr: 2600:1401:2::43 }, addrlen: 28) = -1 ENETUNREACH (Network is unreachable)
0.164 isc-net-0002/1203 sendto(fd: 3, buff: 0x7f73ac064300, len: 107, flags: NOSIGNAL, addr: { .family: UNSPEC }, addr_len: NULL) = 107
0.247 isc-net-0002/1203 connect(fd: 522, uservaddr: { .family: INET6, port: 53, addr: 2600:1401:1::43 }, addrlen: 28) = -1 ENETUNREACH (Network is unreachable)
#
Signed-off-by: Ian Rogers <[email protected]>
Cc: Alexander Shishkin <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Leo Yan <[email protected]>
Cc: Mark Rutland <[email protected]>
Cc: Namhyung Kim <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Stephane Eranian <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
The 'session_done' variable is written to inside the signal handler of
'perf report' and 'perf script'. Switch its type to avoid undefined
behavior.
Signed-off-by: Ian Rogers <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Alexander Shishkin <[email protected]>
Cc: Alexey Bayduraev <[email protected]>
Cc: German Gomez <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Leo Yan <[email protected]>
Cc: Mark Rutland <[email protected]>
Cc: Namhyung Kim <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Stephane Eranian <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Fix to get the declared file name even if it uses file index 0
in DWARF5, using custom die_get_decl_file() function.
Actually, the DWARF5 standard says file index 0 of the DW_AT_decl_file
is invalid(1), but there is a discussion and maybe this will be updated
[2].
Anyway, clang generates such DWARF5 file for the linux kernel. Thus it
must be handled.
Without this, 'perf probe' returns an error:
$ ./perf probe -k $BIN_PATH/vmlinux -s $SRC_PATH -L vfs_read:10
Debuginfo analysis failed.
Error: Failed to show lines.
With this, it can handle the case correctly:
$ ./perf probe -k $BIN_PATH/vmlinux -s $SRC_PATH -L vfs_read:10
<vfs_read@$SRC_PATH/fs/read_write.c:10>
11 ret = rw_verify_area(READ, file, pos, count);
12 if (ret)
return ret;
[1] DWARF5 specification 2.14 says "The value 0 indicates that no source file has been specified.")
[2] http://wiki.dwarfstd.org/index.php?title=DWARF5_Line_Table_File_Numbers)
Signed-off-by: Masami Hiramatsu <[email protected]>
Acked-by: Namhyung Kim <[email protected]>
Cc: Alexander Shishkin <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Mark Rutland <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Steven Rostedt (VMware) <[email protected]>
Link: https://lore.kernel.org/r/166731052936.2100653.13380621874859467731.stgit@devnote3
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Use dwarf_attr_integrate() instead of dwarf_attr() for generic attribute
acccessor functions, so that it can find the specified attribute from
abstact origin DIE etc.
Signed-off-by: Masami Hiramatsu <[email protected]>
Acked-by: Namhyung Kim <[email protected]>
Cc: Alexander Shishkin <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Mark Rutland <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Steven Rostedt (VMware) <[email protected]>
Link: https://lore.kernel.org/r/166731051988.2100653.13595339994343449770.stgit@devnote3
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Since clang generates DWARF5 which sets DW_AT_decl_file as 0,
dwarf_decl_file() thinks that is invalid and returns NULL. In that case
'perf probe' SIGSEGVs because it doesn't expect a NULL decl_file.
This adds a dwarf_decl_file() return value check to avoid such SEGV with
clang generated DWARF5 info.
Without this, 'perf probe' crashes:
$ perf probe -k $BIN_PATH/vmlinux -s $SRC_PATH -L vfs_read:10
Segmentation fault
$
With this, it just warns about it:
$ perf probe -k $BIN_PATH/vmlinux -s $SRC_PATH -L vfs_read:10
Debuginfo analysis failed.
Error: Failed to show lines.
$
Signed-off-by: Masami Hiramatsu <[email protected]>
Acked-by: Namhyung Kim <[email protected]>
Cc: Alexander Shishkin <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Mark Rutland <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Steven Rostedt (VMware) <[email protected]>
Link: https://lore.kernel.org/r/166731051077.2100653.15626653369345128302.stgit@devnote3
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
In most configurations, it works well with skipping 4 entries by
default. If some systems still have 3 BPF internal stack frames,
the next frame should be in a lock function which will be skipped
later when it tries to find a caller. So increasing to 4 won't
affect such systems too.
With --stack-skip=0, I can see something like this:
24 49.84 us 7.41 us 2.08 us mutex bpf_prog_e1b85959d520446c_contention_begin+0x12e
0xffffffffc045040e bpf_prog_e1b85959d520446c_contention_begin+0x12e
0xffffffffc045040e bpf_prog_e1b85959d520446c_contention_begin+0x12e
0xffffffff82ea2071 bpf_trace_run2+0x51
0xffffffff82de775b __bpf_trace_contention_begin+0xb
0xffffffff82c02045 __mutex_lock+0x245
0xffffffff82c019e3 __mutex_lock_slowpath+0x13
0xffffffff82c019c0 mutex_lock+0x20
0xffffffff830a083c kernfs_iop_permission+0x2c
Signed-off-by: Namhyung Kim <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Ian Rogers <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Song Liu <[email protected]>
Cc: [email protected]
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
The msan also warns about the use of VLA for stack_trace variable.
We can dynamically allocate instead. While at it, simplify the error
handle a bit (and fix bugs).
Signed-off-by: Namhyung Kim <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Ian Rogers <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
The msan reported a use-of-uninitialized-value warning for the struct
lock_contention_data in lock_contention_read(). While it'd be filled
by bpf_map_lookup_elem(), let's just initialize it to silence the
warning.
==12524==WARNING: MemorySanitizer: use-of-uninitialized-value
#0 0x562b0f16b1cd in lock_contention_read util/bpf_lock_contention.c:139:7
#1 0x562b0ef65ec6 in __cmd_contention builtin-lock.c:1737:3
#2 0x562b0ef65ec6 in cmd_lock builtin-lock.c:1992:8
#3 0x562b0ee7f50b in run_builtin perf.c:322:11
#4 0x562b0ee7efc1 in handle_internal_command perf.c:376:8
#5 0x562b0ee7e1e9 in run_argv perf.c:420:2
#6 0x562b0ee7e1e9 in main perf.c:550:3
#7 0x7f065f10e632 in __libc_start_main (/usr/lib64/libc.so.6+0x61632)
#8 0x562b0edf2fa9 in _start (perf+0xfa9)
SUMMARY: MemorySanitizer: use-of-uninitialized-value (perf+0xe15160) in lock_contention_read
Signed-off-by: Namhyung Kim <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Ian Rogers <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Needed to get the event_attr_init() and perf_event_paranoid() prototypes
that were being obtained indirectly, by sheer luck.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Those headers are not needed in util/mmap.h, remove them.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Its a thread method, so move it to thread.h, this way some places that
were using event.h just to get this prototype may stop doing so and
speed up building and disentanble the header dependency graph.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Its a addr_location method, so move it to symbol.h, where 'struct
addr_location' is, this way some places that were using event.h just to
get this prototype may stop doing so and speed up building and
disentanble the header dependency graph.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Its a machine method, so move it to machine.h, this way some places that
were using event.h just to get this prototype may stop doing so and
speed up building and disentanble the header dependency graph.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Leave just some forward declarations for pointers, move the includes to
where they are really needed.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
disentangle headers
Some places were including event.h just to get 'struct perf_sample',
move it to a separate place so that we speed up a bit the build.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
map_symbol.h is needed because we have structs that contains 'struct
addr_map_symbol', so add it, remove the others.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
In the bpf-prologue.h header we are just using pointers, so no need to
include headers for that, just provide forward declarations for those
types.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Use the global quiet variable everywhere so that all tools hide warnings
in quiet mode and update the documentation to reflect this.
'perf probe' claimed that errors are not printed in quiet mode but I
don't see this so remove it from the docs.
Signed-off-by: James Clark <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Alexander Shishkin <[email protected]>
Cc: Ian Rogers <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Mark Rutland <[email protected]>
Cc: Namhyung Kim <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Since commit c897899752478d4c ("perf tools: Prevent out-of-bounds access
to registers") the util/event.h header doesn't use anything from
util/perf_regs.h, so drop it to untangle the header dependency tree a
bit, speeding up compilation.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
perf_reg_name() prototype
It was getting it via event.h, that doesn't need that include anymore
and will drop it.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
To reduce compile time and header dependency chains just add forward
declarations for pointer types and include linux/types.h for u64.
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Add missing newline on pr_warning() call in home_perfconfig().
Before:
# perf record
File /home/yangjihong/.perfconfig not owned by current user or root, ignoring it.Couldn't synthesize bpf events.
After:
# perf record
File /home/yangjihong/.perfconfig not owned by current user or root, ignoring it.
Couldn't synthesize bpf events.
Signed-off-by: Yang Jihong <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Alexander Shishkin <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Mark Rutland <[email protected]>
Cc: Namhyung Kim <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
The aggr field in the struct perf_counts is to keep the aggregated value
in the AGGR_GLOBAL for the old code. But it's not used anymore.
Signed-off-by: Namhyung Kim <[email protected]>
Acked-by: Ian Rogers <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Athira Jajeev <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Leo Yan <[email protected]>
Cc: Michael Petlan <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
The recent change in the perf stat broke the percore event display.
Note that the aggr counts are already processed so that the every
sibling thread in the same core will get the per-core counter values.
Check percore evsels and skip the sibling threads in the display.
Signed-off-by: Namhyung Kim <[email protected]>
Acked-by: Ian Rogers <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Athira Jajeev <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Leo Yan <[email protected]>
Cc: Michael Petlan <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Now aggr counts are ready for use. Convert the display routines to use
the aggr counts and update the shadow stat with them. It doesn't need
to aggregate counts or collect aliases anymore during the display. Get
rid of now unused struct perf_aggr_thread_value.
Note that there's a difference in the display order among the aggr mode.
For per-core/die/socket/node aggregation, it shows relevant events in
the same unit together, whereas global/thread/no aggregation it shows
the same events for different units together. So it still uses separate
codes to display them due to the ordering.
One more thing to note is that it breaks per-core event display for now.
The next patch will fix it to have identical output as of now.
Signed-off-by: Namhyung Kim <[email protected]>
Acked-by: Ian Rogers <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Athira Jajeev <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Leo Yan <[email protected]>
Cc: Michael Petlan <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
This function updates the shadow stats using the aggregated counts
uniformly since it uses the aggr_counts for the every aggr mode.
It'd have duplicate shadow stats for each items for now since the
display routines will update them once again. But that'd be fine
as it shows the average values and it'd be gone eventually.
Signed-off-by: Namhyung Kim <[email protected]>
Acked-by: Ian Rogers <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Athira Jajeev <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Leo Yan <[email protected]>
Cc: Michael Petlan <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
The perf_stat_process_percore() is to aggregate counts for an event per-core
even if the aggr_mode is AGGR_NONE. This is enabled when user requested it
on the command line.
To handle that, it keeps the per-cpu counts at first. And then it aggregates
the counts that have the same core id in the aggr->counts and updates the
values for each cpu back.
Later, per-core events will skip one of the CPUs unless percore-show-thread
option is given. In that case, it can simply print all cpu stats with the
updated (per-core) values.
Signed-off-by: Namhyung Kim <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Athira Jajeev <[email protected]>
Cc: Ian Rogers <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Leo Yan <[email protected]>
Cc: Michael Petlan <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
The perf_stat_merge_counters() is to aggregate the same events in different
PMUs like in case of uncore or hybrid. The same logic is in the stat-display
routines but I think it should be handled when it processes the event counters.
As it works on the aggr_counters, it doesn't change the output yet.
Signed-off-by: Namhyung Kim <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Athira Jajeev <[email protected]>
Cc: Ian Rogers <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Leo Yan <[email protected]>
Cc: Michael Petlan <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
The evsel->stats->aggr->count should be reset for interval processing
since we want to use the values directly for display.
Signed-off-by: Namhyung Kim <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Athira Jajeev <[email protected]>
Cc: Ian Rogers <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Leo Yan <[email protected]>
Cc: Michael Petlan <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
In the process_stat_config_event() it sets the aggr_mode that means the
earlier evlist__alloc_stats() cannot allocate the aggr counts due to the
missing aggr_mode.
Do it after setting the aggr_map using evlist__alloc_aggr_stats().
Signed-off-by: Namhyung Kim <[email protected]>
Acked-by: Ian Rogers <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Athira Jajeev <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Leo Yan <[email protected]>
Cc: Michael Petlan <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Per-thread aggregation doesn't use the CPU numbers but the logic should
be the same. Initialize cpu_aggr_map separately for AGGR_THREAD and use
thread map idx to aggregate counter values.
Signed-off-by: Namhyung Kim <[email protected]>
Acked-by: Ian Rogers <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Athira Jajeev <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Leo Yan <[email protected]>
Cc: Michael Petlan <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
It's possible to have 0 enabled/running time for some per-task or per-cgroup
events since it's not scheduled on any CPU. Treating the whole event as
failed would not work in this case. Thinking again, the code only existed
when any CPU-level aggregation is enabled (like per-socket, per-core, ...).
To make it clearer, factor out the condition check into the new
evsel__count_has_error() function and add some comments.
Signed-off-by: Namhyung Kim <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Athira Jajeev <[email protected]>
Cc: Ian Rogers <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Leo Yan <[email protected]>
Cc: Michael Petlan <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|
|
Add a logic to aggregate counter values to the new evsel->stats->aggr.
This is not used yet so shadow stats are not updated. But later patch
will convert the existing code to use it.
With that, we don't need to handle AGGR_GLOBAL specially anymore. It
can use the same logic with counts, prev_counts and aggr_counts.
Signed-off-by: Namhyung Kim <[email protected]>
Cc: Adrian Hunter <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Athira Jajeev <[email protected]>
Cc: Ian Rogers <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: James Clark <[email protected]>
Cc: Jiri Olsa <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Leo Yan <[email protected]>
Cc: Michael Petlan <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Xing Zhengjun <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Arnaldo Carvalho de Melo <[email protected]>
|