linux_dsm_epyc7002/tools/perf
Borislav Petkov 02d492e5dc perf stat: Issue a HW watchdog disable hint
When using perf stat on an AMD F15h system with the default hw events
attributes, some of the events don't get counted:

 Performance counter stats for 'sleep 1':

          0.749208      task-clock (msec)         #    0.001 CPUs utilized
                 1      context-switches          #    0.001 M/sec
                 0      cpu-migrations            #    0.000 K/sec
                54      page-faults               #    0.072 M/sec
         1,122,815      cycles                    #    1.499 GHz
           286,740      stalled-cycles-frontend   #   25.54% frontend cycles idle
     <not counted>      stalled-cycles-backend                                        (0.00%)
     ^^^^^^^^^^^^
     <not counted>      instructions                                                  (0.00%)
     ^^^^^^^^^^^^
     <not counted>      branches                                                      (0.00%)
     <not counted>      branch-misses                                                 (0.00%)

       1.001550070 seconds time elapsed

The reason is that we have the HW watchdog consuming one PMU counter and
when perf tries to schedule 6 events on 6 counters and some of those
counters are constrained to only a specific subset of PMCs by the
hardware, the event scheduling fails.

So issue a hint to disable the HW watchdog around a perf stat session.

Committer note:

Testing it...

  # perf stat -d usleep 1

   Performance counter stats for 'usleep 1':

          1.180203      task-clock (msec)         #    0.490 CPUs utilized
                 1      context-switches          #    0.847 K/sec
                 0      cpu-migrations            #    0.000 K/sec
                54      page-faults               #    0.046 M/sec
           184,754      cycles                    #    0.157 GHz
           714,553      instructions              #    3.87  insn per cycle
           154,661      branches                  #  131.046 M/sec
             7,247      branch-misses             #    4.69% of all branches
           219,984      L1-dcache-loads           #  186.395 M/sec
            17,600      L1-dcache-load-misses     #    8.00% of all L1-dcache hits    (90.16%)
     <not counted>      LLC-loads                                                     (0.00%)
     <not counted>      LLC-load-misses                                               (0.00%)

       0.002406823 seconds time elapsed

  Some events weren't counted. Try disabling the NMI watchdog:
	echo 0 > /proc/sys/kernel/nmi_watchdog
	perf stat ...
	echo 1 > /proc/sys/kernel/nmi_watchdog
  #

Signed-off-by: Borislav Petkov <bp@suse.de>
Acked-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Robert Richter <rric@kernel.org>
Cc: Vince Weaver <vince@deater.net>
Link: http://lkml.kernel.org/r/20170211183218.ijnvb5f7ciyuunx4@pd.tnic
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2017-03-03 19:07:13 -03:00
..
arch perf tools arm64: Add support for generating bpf prologue 2017-02-08 08:55:01 -03:00
bench perf bench numa: Make sure dprintf() is not defined 2017-02-14 15:19:18 -03:00
Documentation Merge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2017-02-28 11:38:18 -08:00
jvmti perf kvmti: Remove unused Makefile file 2016-11-14 12:42:56 -03:00
pmu-events perf vendor events: Add mapping for KnightsMill PMU events 2017-03-03 19:07:13 -03:00
python perf python: Add tracepoint example 2016-07-12 16:23:35 -03:00
scripts perf/core improvements and fixes: 2016-08-04 11:02:38 +02:00
tests perf utils: Check verbose flag properly 2017-02-20 11:35:54 -03:00
trace perf trace: Check if MAP_32BIT is defined (again) 2016-12-20 09:37:40 -03:00
ui perf utils: Check verbose flag properly 2017-02-20 11:35:54 -03:00
util Merge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2017-02-28 11:38:18 -08:00
.gitignore perf tools: Add arch/*/include/generated/ to .gitignore 2016-05-30 12:41:46 -03:00
Build perf ftrace: Introduce new 'ftrace' tool 2017-01-26 11:43:01 -03:00
builtin-annotate.c perf annotate: Add -q/--quiet option 2017-02-20 11:47:18 -03:00
builtin-bench.c perf subcmd: Create subcmd library 2015-12-17 14:27:14 -03:00
builtin-buildid-cache.c tools: Introduce str_error_r() 2016-07-12 15:19:47 -03:00
builtin-buildid-list.c perf subcmd: Create subcmd library 2015-12-17 14:27:14 -03:00
builtin-c2c.c perf c2c report: Coalesce by default only by pid,iaddr 2017-01-20 16:52:56 -03:00
builtin-config.c perf config: Mark where are config items from (user or system) 2016-11-14 13:10:37 -03:00
builtin-data.c perf data ctf: Add '--all' option for 'perf data convert' 2016-06-28 10:54:57 -03:00
builtin-diff.c perf diff: Add -q/--quiet option 2017-02-20 11:47:08 -03:00
builtin-evlist.c perf evlist: Rename for_each() macros to for_each_entry() 2016-06-23 11:26:15 -03:00
builtin-ftrace.c perf ftrace: Add ftrace.tracer config option 2017-01-31 16:20:09 -03:00
builtin-help.c perf tools: Propagate perf_config() errors 2017-01-27 12:23:33 -03:00
builtin-inject.c perf symbols: Remove symbol_filter_t machinery 2016-09-05 11:14:50 -03:00
builtin-kallsyms.c perf kallsyms: Introduce tool to look for extended symbol information on the running kernel 2017-01-11 16:48:01 -03:00
builtin-kmem.c perf symbols: No need to check if sym->name is NULL 2017-02-13 17:22:34 -03:00
builtin-kvm.c perf kvm: Use NSEC_PER_USEC 2016-08-23 15:37:33 -03:00
builtin-list.c perf list: Add debug support for outputing alias string 2017-02-08 08:55:04 -03:00
builtin-lock.c perf subcmd: Create subcmd library 2015-12-17 14:27:14 -03:00
builtin-mem.c perf utils: Check verbose flag properly 2017-02-20 11:35:54 -03:00
builtin-probe.c perf probe: Add option --symfs 2017-02-08 09:41:10 -03:00
builtin-record.c perf record: Honor --quiet option properly 2017-02-20 11:50:36 -03:00
builtin-report.c perf report: Add -q/--quiet option 2017-02-20 11:46:58 -03:00
builtin-sched.c perf utils: Check verbose flag properly 2017-02-20 11:35:54 -03:00
builtin-script.c perf script: Also allow forcing reading of non-root owned files by root 2017-01-16 14:59:15 -03:00
builtin-stat.c perf stat: Issue a HW watchdog disable hint 2017-03-03 19:07:13 -03:00
builtin-timechart.c perf timechart: Use NSEC_PER_U?SEC 2016-08-23 15:37:33 -03:00
builtin-top.c perf utils: Check verbose flag properly 2017-02-20 11:35:54 -03:00
builtin-trace.c perf utils: Check verbose flag properly 2017-02-20 11:35:54 -03:00
builtin-version.c perf tools: Move cmd_version() to builtin-version.c 2015-12-09 13:42:03 -03:00
builtin.h perf ftrace: Introduce new 'ftrace' tool 2017-01-26 11:43:01 -03:00
check-headers.sh perf tools: Move headers check into bash script 2016-12-15 16:25:44 -03:00
command-list.txt perf ftrace: Introduce new 'ftrace' tool 2017-01-26 11:43:01 -03:00
CREDITS
design.txt
Makefile perf build tests: Do parallell builds with 'build-test' 2016-02-04 15:57:00 -03:00
Makefile.config tools perf scripting python: clang doesn't have -spec, remove it 2017-02-17 10:31:13 -03:00
Makefile.perf perf build: Add special fixdep cleaning rule 2017-02-17 16:04:38 -03:00
MANIFEST tools include: Introduce linux/compiler-gcc.h 2017-02-13 14:29:29 -03:00
perf-archive.sh
perf-completion.sh perf tools: Avoid confusion with preloaded bash function for perf bash completion 2015-03-19 13:53:27 -03:00
perf-read-vdso.c
perf-sys.h perf powerpc: Fix build-test failure 2016-09-08 13:44:07 -03:00
perf-with-kcore.sh perf tools: Fix perf-with-kcore handling of arguments containing spaces 2015-08-06 16:48:27 -03:00
perf.c perf tools: Propagate perf_config() errors 2017-01-27 12:23:33 -03:00
perf.h perf evsel: Allow to ignore missing pid 2016-12-15 16:25:46 -03:00