linux-stable/arch/x86/events/intel
Kan Liang f861854e1b perf/x86/intel: Fix inaccurate period in context switch for auto-reload
Perf doesn't take the left period into account when auto-reload is
enabled with fixed period sampling mode in context switch.

Here is the MSR trace of the perf command as below.
(The MSR trace is simplified from a ftrace log.)

    #perf record -e cycles:p -c 2000000 -- ./triad_loop

      //The MSR trace of task schedule out
      //perf disable all counters, disable PEBS, disable GP counter 0,
      //read GP counter 0, and re-enable all counters.
      //The counter 0 stops at 0xfffffff82840
      write_msr: MSR_CORE_PERF_GLOBAL_CTRL(38f), value 0
      write_msr: MSR_IA32_PEBS_ENABLE(3f1), value 0
      write_msr: MSR_P6_EVNTSEL0(186), value 40003003c
      rdpmc: 0, value fffffff82840
      write_msr: MSR_CORE_PERF_GLOBAL_CTRL(38f), value f000000ff

      //The MSR trace of the same task schedule in again
      //perf disable all counters, enable and set GP counter 0,
      //enable PEBS, and re-enable all counters.
      //0xffffffe17b80 (-2000000) is written to GP counter 0.
      write_msr: MSR_CORE_PERF_GLOBAL_CTRL(38f), value 0
      write_msr: MSR_IA32_PMC0(4c1), value ffffffe17b80
      write_msr: MSR_P6_EVNTSEL0(186), value 40043003c
      write_msr: MSR_IA32_PEBS_ENABLE(3f1), value 1
      write_msr: MSR_CORE_PERF_GLOBAL_CTRL(38f), value f000000ff

When the same task schedule in again, the counter should starts from
previous left. However, it starts from the fixed period -2000000 again.

A special variant of intel_pmu_save_and_restart() is used for
auto-reload, which doesn't update the hwc->period_left.
When the monitored task schedules in again, perf doesn't know the left
period. The fixed period is used, which is inaccurate.

With auto-reload, the counter always has a negative counter value. So
the left period is -value. Update the period_left in
intel_pmu_save_and_restart_reload().

With the patch:

      //The MSR trace of task schedule out
      write_msr: MSR_CORE_PERF_GLOBAL_CTRL(38f), value 0
      write_msr: MSR_IA32_PEBS_ENABLE(3f1), value 0
      write_msr: MSR_P6_EVNTSEL0(186), value 40003003c
      rdpmc: 0, value ffffffe25cbc
      write_msr: MSR_CORE_PERF_GLOBAL_CTRL(38f), value f000000ff

      //The MSR trace of the same task schedule in again
      write_msr: MSR_CORE_PERF_GLOBAL_CTRL(38f), value 0
      write_msr: MSR_IA32_PMC0(4c1), value ffffffe25cbc
      write_msr: MSR_P6_EVNTSEL0(186), value 40043003c
      write_msr: MSR_IA32_PEBS_ENABLE(3f1), value 1
      write_msr: MSR_CORE_PERF_GLOBAL_CTRL(38f), value f000000ff

Fixes: d31fc13fdc ("perf/x86/intel: Fix event update for auto-reload")
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lkml.kernel.org/r/20200121190125.3389-1-kan.liang@linux.intel.com
2020-02-11 13:23:27 +01:00
..
bts.c perf/x86/intel/bts: Fix the use of page_private() 2019-12-17 13:32:46 +01:00
core.c perf/x86/intel: Add Elkhart Lake support 2020-02-11 13:17:48 +01:00
cstate.c perf/x86/cstate: Add Tremont support 2020-02-11 13:17:49 +01:00
ds.c perf/x86/intel: Fix inaccurate period in context switch for auto-reload 2020-02-11 13:23:27 +01:00
knc.c License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
lbr.c perf/x86/intel: Implement LBR callstack context synchronization 2019-10-28 12:51:01 +01:00
Makefile License cleanup: add SPDX GPL-2.0 license identifier to files with no license 2017-11-02 11:10:55 +01:00
p4.c perf_event: Add support for LSM and SELinux checks 2019-10-17 21:31:55 +02:00
p6.c x86/cpu: Rename cpu_data.x86_mask to cpu_data.x86_stepping 2018-02-15 01:15:52 +01:00
pt.c perf/x86/intel/pt: Prevent redundant WRMSRs 2019-11-13 11:06:18 +01:00
pt.h perf/x86/intel/pt: Prevent redundant WRMSRs 2019-11-13 11:06:18 +01:00
rapl.c perf/x86/intel/rapl: Add Comet Lake support 2020-01-17 10:19:25 +01:00
uncore.c perf/x86/uncore: Fix event group support 2019-10-28 11:02:01 +01:00
uncore.h perf/x86/uncore: Fix event group support 2019-10-28 11:02:01 +01:00
uncore_nhmex.c perf/x86/intel/uncore: Correct fixed counter index check for NHM 2018-05-31 12:36:28 +02:00
uncore_snb.c perf/x86/intel/uncore: Add PCI ID of IMC for Xeon E3 V5 Family 2020-01-17 11:33:18 +01:00
uncore_snbep.c perf/x86/intel/uncore: Remove PCIe3 unit for SNR 2020-01-17 11:33:38 +01:00