LinuxLists.cc - [PATCH] irqflags: Remove the CONFIG_TRACE

2021-05-24 02:52:30

Subject: [PATCH] irqflags: Remove the CONFIG_TRACE_IRQFLAGS

trace_hardirqs_on() and trace_hardirqs_off() had include the
CONFIG_TRACE_IRQFLAGS, so remove the redundant CONFIG_TRACE_IRQFLAGS.

Signed-off-by: Yejune Deng <[email protected]>
---
include/linux/irqflags.h | 16 ----------------
1 file changed, 16 deletions(-)

diff --git a/include/linux/irqflags.h b/include/linux/irqflags.h
index 600c10d..b5edd7b 100644
--- a/include/linux/irqflags.h
+++ b/include/linux/irqflags.h
@@ -189,12 +189,6 @@ extern void warn_bogus_irq_restore(void);
#define raw_irqs_disabled() (arch_irqs_disabled())
#define raw_safe_halt() arch_safe_halt()

-/*
- * The local_irq_*() APIs are equal to the raw_local_irq*()
- * if !TRACE_IRQFLAGS.
- */
-#ifdef CONFIG_TRACE_IRQFLAGS
-
#define local_irq_enable() \
do { \
trace_hardirqs_on(); \
@@ -230,16 +224,6 @@ extern void warn_bogus_irq_restore(void);
} while (0)

-#else /* !CONFIG_TRACE_IRQFLAGS */
-
-#define local_irq_enable() do { raw_local_irq_enable(); } while (0)
-#define local_irq_disable() do { raw_local_irq_disable(); } while (0)
-#define local_irq_save(flags) do { raw_local_irq_save(flags); } while (0)
-#define local_irq_restore(flags) do { raw_local_irq_restore(flags); } while (0)
-#define safe_halt() do { raw_safe_halt(); } while (0)
-
-#endif /* CONFIG_TRACE_IRQFLAGS */
-
#define local_save_flags(flags) raw_local_save_flags(flags)

/*
--
2.7.4

2021-05-24 11:29:58

by Mark Rutland

[permalink] [raw]

Subject: Re: [PATCH] irqflags: Remove the CONFIG_TRACE_IRQFLAGS

On Mon, May 24, 2021 at 10:51:19AM +0800, Yejune Deng wrote:
> trace_hardirqs_on() and trace_hardirqs_off() had include the
> CONFIG_TRACE_IRQFLAGS, so remove the redundant CONFIG_TRACE_IRQFLAGS.
>
> Signed-off-by: Yejune Deng <[email protected]>
> ---
> include/linux/irqflags.h | 16 ----------------
> 1 file changed, 16 deletions(-)
>
> diff --git a/include/linux/irqflags.h b/include/linux/irqflags.h
> index 600c10d..b5edd7b 100644
> --- a/include/linux/irqflags.h
> +++ b/include/linux/irqflags.h
> @@ -189,12 +189,6 @@ extern void warn_bogus_irq_restore(void);
> #define raw_irqs_disabled() (arch_irqs_disabled())
> #define raw_safe_halt() arch_safe_halt()
>
> -/*
> - * The local_irq_*() APIs are equal to the raw_local_irq*()
> - * if !TRACE_IRQFLAGS.
> - */
> -#ifdef CONFIG_TRACE_IRQFLAGS
> -
> #define local_irq_enable() \
> do { \
> trace_hardirqs_on(); \
> @@ -230,16 +224,6 @@ extern void warn_bogus_irq_restore(void);
> } while (0)
>
>
> -#else /* !CONFIG_TRACE_IRQFLAGS */
> -
> -#define local_irq_enable() do { raw_local_irq_enable(); } while (0)
> -#define local_irq_disable() do { raw_local_irq_disable(); } while (0)

The implementation we're left with is:

| #define local_irq_disable() \
| do { \
| bool was_disabled = raw_irqs_disabled();\
| raw_local_irq_disable(); \
| if (!was_disabled) \
| trace_hardirqs_off(); \
| } while (0)

... where the compiler almost certainly won't be able to optimize away
the call to raw_irqs_disabled(), which wraps arch_irqs_disabled(). Note
that arch_irqs_disabled() is typically implemented with a volatile asm
block to save the processor flags.

We could special-case local_irq_disable() to prevent this, or we could
leave it as-is. I'd err towards the latter.

Thanks,
Mark.

> -#define local_irq_save(flags) do { raw_local_irq_save(flags); } while (0)
> -#define local_irq_restore(flags) do { raw_local_irq_restore(flags); } while (0)
> -#define safe_halt() do { raw_safe_halt(); } while (0)
> -
> -#endif /* CONFIG_TRACE_IRQFLAGS */
> -
> #define local_save_flags(flags) raw_local_save_flags(flags)
>
> /*
> --
> 2.7.4
>

2021-05-25 06:50:00

by kernel test robot

[permalink] [raw]

Subject: [irqflags] 21db66c4ff: will-it-scale.per_process_ops -4.5% regression

Greeting,

FYI, we noticed a

commit: 21db66c4ff4e7378757 url: in testcase: will-it-scale
on test machine: 192 with following parameters:

nr_task: 16
mode: process
test: dup1
cpufreq_governor: performance
ucode: 0x5003006

test-description: test-url: Thanks,
Oliver Sang

-4.5% regression of will-it-scale.per_process_ops due to commit:
27ee7a1af94fd30c2a2f4 ("[PATCH] irqflags: Remove the CONFIG_TRACE_IRQFLAGS")
ub.com/0day-ci/linux/commits/Yejune-Deng/irqflags-Remove-the-CONFIG_TRACE_IRQFLAGS/20210524-105251">https://github.com/0day-ci/linux/commits/Yejune-Deng/irqflags-Remove-the-CONFIG_TRACE_IRQFLAGS/20210524-105251
.kernel.org/cgit/linux/kernel/git/torvalds/linux.git">https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git c4681547bcce777daf576925a966ffa824edd09d
threads 4 sockets Intel(R) Xeon(R) Platinum 9242 CPU @ 2.30GHz with 192G memory
Will It Scale takes a testcase and runs it from 1 through to n parallel copies to see if the testcase will scale. It builds both a process and threads based test in order to see any differences between the two.
/github.com/antonblanchard/will-it-scale">https://github.com/antonblanchard/will-it-scale
the commit also has significant impact on the following tests:
--------------------------------------------------------+
| will-it-scale: will-it-scale.per_process_ops -1.2% regression |
| 104 threads 2 sockets Skylake with 192G memory |
| cpufreq_governor=performance |
|
|
|
|
--------------------------------------------------------+
kconfig for both 21db66c4ff and parent (config file is attached)
is disabled.
kindly add following tag
test robot <[email protected]>
----------------------------------------------------------------------->
href="https://github.com/intel/lkp-tests.git">https://github.com/intel/lkp-tests.git
job.yaml # job file is attached in this email
--compatible job.yaml # generate the yaml file for lkp run
generated-yaml-file
==============================================================
config/mode/nr_task/rootfs/tbox_group/test/testcase/ucode:
rhel-8.3/process/16/debian-10.4-x86_64-20200603.cgz/lkp-csl-2ap2/dup1/will-it-scale/0x5003006
Remove the CONFIG_TRACE_IRQFLAGS")
4e737875727ee7a1a
-----------------
%stddev
\
-4.5% 74526776 will-it-scale.16.processes
-4.5% 4657922 will-it-scale.per_process_ops
-4.5% 74526776 will-it-scale.workload
50161 ? 2% meminfo.Active
50161 ? 2% meminfo.Active(anon)
26213 ? 14% softirqs.CPU104.SCHED
24899 ? 10% softirqs.CPU109.SCHED
12576 ? 17% softirqs.CPU162.RCU
-46.6% 2.692e+10 ? 78% cpuidle.C1E.time
-29.7% 73111874 ? 29% cpuidle.C1E.usage
+1300.9% 2.459e+10 ? 85% cpuidle.C6.time
+689.3% 29200515 ? 84% cpuidle.C6.usage
0.05 ? 67% perf-sched.sch_delay.max.ms.do_task_dead.do_exit.do_group_exit.__x64_sys_exit_group.do_syscall_64
0.08 ? 85% perf-sched.sch_delay.max.ms.do_wait.kernel_wait4.__do_sys_wait4.do_syscall_64
681.63 ? 5% perf-sched.wait_and_delay.avg.ms.schedule_hrtimeout_range_clock.poll_schedule_timeout.constprop.0.do_sys_poll
50.14 ? 4% perf-sched.wait_and_delay.count.schedule_hrtimeout_range_clock.poll_schedule_timeout.constprop.0.do_sys_poll
681.63 ? 5% perf-sched.wait_time.avg.ms.schedule_hrtimeout_range_clock.poll_schedule_timeout.constprop.0.do_sys_poll
12517 ? 2% proc-vmstat.nr_active_anon
17532 proc-vmstat.nr_shmem
12517 ? 2% proc-vmstat.nr_zone_active_anon
34224 ? 39% proc-vmstat.numa_pte_updates
16757 proc-vmstat.pgactivate
24.00 ? 27% turbostat.Avg_MHz
-31.7% 39572049 ? 32% turbostat.C1E
24.59 ? 86% turbostat.C1E%
+693.9% 16821659 ? 85% turbostat.C6
24.54 ? 85% turbostat.C6%
16.01 ? 88% turbostat.CPU%c6
1.17 ? 77% perf-stat.i.MPKI
-4.3% 1.164e+10 perf-stat.i.branch-instructions
6.41 ? 49% perf-stat.i.cache-miss-rate%
+914.4% 70145425 ? 77% perf-stat.i.cache-references
0.93 ? 3% perf-stat.i.cpi
+4.6% 5.595e+10 ? 3% perf-stat.i.cpu-cycles
0.01 ? 85% perf-stat.i.dTLB-load-miss-rate%
-3.1% 1.941e+10 perf-stat.i.dTLB-loads
0.00 ? 77% perf-stat.i.dTLB-store-miss-rate%
281623 ? 77% perf-stat.i.dTLB-store-misses
-3.3% 1.302e+10 perf-stat.i.dTLB-stores
-3.9% 6.041e+10 perf-stat.i.instructions
1.08 ? 3% perf-stat.i.ipc
0.29 ? 3% perf-stat.i.metric.GHz
379.32 ? 74% perf-stat.i.metric.K/sec
-3.5% 229.53 perf-stat.i.metric.M/sec
-9.6% 171895 ? 7% perf-stat.i.node-load-misses
77.65 ? 18% perf-stat.i.node-store-miss-rate%
1.16 ? 77% perf-stat.overall.MPKI
6.53 ? 50% perf-stat.overall.cache-miss-rate%
0.93 ? 3% perf-stat.overall.cpi
0.01 ? 85% perf-stat.overall.dTLB-load-miss-rate%
0.00 ? 77% perf-stat.overall.dTLB-store-miss-rate%
1.08 ? 3% perf-stat.overall.ipc
76.04 ? 16% perf-stat.overall.node-store-miss-rate%
-4.3% 1.16e+10 perf-stat.ps.branch-instructions
+914.6% 69905237 ? 77% perf-stat.ps.cache-references
+4.6% 5.576e+10 ? 3% perf-stat.ps.cpu-cycles
2158934 ? 85% perf-stat.ps.dTLB-load-misses
-3.1% 1.934e+10 perf-stat.ps.dTLB-loads
280650 ? 77% perf-stat.ps.dTLB-store-misses
-3.3% 1.298e+10 perf-stat.ps.dTLB-stores
-3.9% 6.021e+10 perf-stat.ps.instructions
-9.7% 171303 ? 7% perf-stat.ps.node-load-misses
-3.9% 1.818e+13 perf-stat.total.instructions
0.86 ? 7% perf-profile.calltrace.cycles-pp.locks_remove_posix.filp_close.__x64_sys_close.do_syscall_64.entry_SYSCALL_64_after_hwframe
1.07 ? 5% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_safe_stack.__close
0.68 ? 10% perf-profile.calltrace.cycles-pp.dnotify_flush.filp_close.__x64_sys_close.do_syscall_64.entry_SYSCALL_64_after_hwframe
1.63 ? 7% perf-profile.calltrace.cycles-pp.hrtimer_interrupt.__sysvec_apic_timer_interrupt.sysvec_apic_timer_interrupt.asm_sysvec_apic_timer_interrupt.cpuidle_enter_state
0.87 ? 16% perf-profile.calltrace.cycles-pp.menu_select.do_idle.cpu_startup_entry.start_secondary.secondary_startup_64_no_verify
0.98 ? 16% perf-profile.calltrace.cycles-pp.__hrtimer_run_queues.hrtimer_interrupt.__sysvec_apic_timer_interrupt.sysvec_apic_timer_interrupt.asm_sysvec_apic_timer_interrupt
1.66 ? 7% perf-profile.calltrace.cycles-pp.__sysvec_apic_timer_interrupt.sysvec_apic_timer_interrupt.asm_sysvec_apic_timer_interrupt.cpuidle_enter_state.cpuidle_enter
2.31 ? 3% perf-profile.calltrace.cycles-pp.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.dup
2.55 ? 7% perf-profile.calltrace.cycles-pp.sysvec_apic_timer_interrupt.asm_sysvec_apic_timer_interrupt.cpuidle_enter_state.cpuidle_enter.do_idle
2.85 ? 7% perf-profile.calltrace.cycles-pp.asm_sysvec_apic_timer_interrupt.cpuidle_enter_state.cpuidle_enter.do_idle.cpu_startup_entry
2.47 ? 5% perf-profile.calltrace.cycles-pp.exit_to_user_mode_prepare.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.__close
4.14 ? 6% perf-profile.calltrace.cycles-pp.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.__close
15.49 ? 5% perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe.__close
17.50 ? 5% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe.__close
0.12 ? 10% perf-profile.children.cycles-pp.read_tsc
0.15 ? 9% perf-profile.children.cycles-pp.process_simple
0.09 ? 23% perf-profile.children.cycles-pp.update_rq_clock
0.06 ? 7% perf-profile.children.cycles-pp.perf_session__deliver_event
0.15 ? 7% perf-profile.children.cycles-pp.native_sched_clock
0.11 ? 15% perf-profile.children.cycles-pp.load_balance
0.07 ? 11% perf-profile.children.cycles-pp.update_sd_lb_stats
0.07 ? 18% perf-profile.children.cycles-pp.hrtimer_next_event_without
0.05 ? 9% perf-profile.children.cycles-pp.__ordered_events__flush
0.15 ? 15% perf-profile.children.cycles-pp.perf_mux_hrtimer_handler
0.19 ? 8% perf-profile.children.cycles-pp.rebalance_domains
0.17 ? 8% perf-profile.children.cycles-pp.sched_clock_cpu
0.06 ? 16% perf-profile.children.cycles-pp.arch_scale_freq_tick
0.08 ? 10% perf-profile.children.cycles-pp.find_busiest_group
0.07 ? 17% perf-profile.children.cycles-pp.get_next_timer_interrupt
0.08 ? 29% perf-profile.children.cycles-pp.rcu_idle_exit
0.11 ? 23% perf-profile.children.cycles-pp.rcu_sched_clock_irq
0.08 ? 31% perf-profile.children.cycles-pp.__hrtimer_next_event_base
0.09 ? 33% perf-profile.children.cycles-pp.hrtimer_update_next_event
0.48 ? 8% perf-profile.children.cycles-pp.rcu_nocb_flush_deferred_wakeup
0.33 ? 11% perf-profile.children.cycles-pp.__softirqentry_text_start
0.43 ? 10% perf-profile.children.cycles-pp.irq_exit_rcu
0.51 ? 7% perf-profile.children.cycles-pp.scheduler_tick
0.90 ? 8% perf-profile.children.cycles-pp.locks_remove_posix
0.68 ? 10% perf-profile.children.cycles-pp.dnotify_flush
0.79 ? 8% perf-profile.children.cycles-pp.update_process_times
0.85 ? 6% perf-profile.children.cycles-pp.syscall_exit_to_user_mode_prepare
0.84 ? 9% perf-profile.children.cycles-pp.tick_sched_handle
0.91 ? 7% perf-profile.children.cycles-pp.tick_sched_timer
0.88 ? 15% perf-profile.children.cycles-pp.menu_select
1.29 ? 11% perf-profile.children.cycles-pp.__hrtimer_run_queues
2.02 ? 3% perf-profile.children.cycles-pp.hrtimer_interrupt
2.06 ? 3% perf-profile.children.cycles-pp.__sysvec_apic_timer_interrupt
2.98 ? 5% perf-profile.children.cycles-pp.sysvec_apic_timer_interrupt
3.32 ? 5% perf-profile.children.cycles-pp.asm_sysvec_apic_timer_interrupt
3.46 ? 4% perf-profile.children.cycles-pp.exit_to_user_mode_prepare
6.91 ? 4% perf-profile.children.cycles-pp.syscall_exit_to_user_mode
33.56 ? 4% perf-profile.children.cycles-pp.do_syscall_64
37.72 ? 4% perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe
0.14 ? 8% perf-profile.self.cycles-pp.native_sched_clock
0.06 ? 16% perf-profile.self.cycles-pp.arch_scale_freq_tick
0.08 ? 57% perf-profile.self.cycles-pp.tick_nohz_next_event
0.08 ? 28% perf-profile.self.cycles-pp.do_idle
0.09 ? 25% perf-profile.self.cycles-pp.rcu_sched_clock_irq
0.37 ? 8% perf-profile.self.cycles-pp.rcu_nocb_flush_deferred_wakeup
0.41 ? 9% perf-profile.self.cycles-pp.cpuidle_enter_state
0.42 ? 10% perf-profile.self.cycles-pp.menu_select
0.67 ? 11% perf-profile.self.cycles-pp.dnotify_flush
0.81 ? 6% perf-profile.self.cycles-pp.syscall_exit_to_user_mode_prepare
2.87 ? 4% perf-profile.self.cycles-pp.exit_to_user_mode_prepare
2.81 ? 4% perf-profile.self.cycles-pp.syscall_exit_to_user_mode
3677 ? 59% interrupts.CPU1.NMI:Non-maskable_interrupts
3677 ? 59% interrupts.CPU1.PMI:Performance_monitoring_interrupts
2650 ? 23% interrupts.CPU101.NMI:Non-maskable_interrupts
2650 ? 23% interrupts.CPU101.PMI:Performance_monitoring_interrupts
3272 ? 41% interrupts.CPU104.NMI:Non-maskable_interrupts
3272 ? 41% interrupts.CPU104.PMI:Performance_monitoring_interrupts
129.43 ? 33% interrupts.CPU120.NMI:Non-maskable_interrupts
129.43 ? 33% interrupts.CPU120.PMI:Performance_monitoring_interrupts
120.71 ? 31% interrupts.CPU137.NMI:Non-maskable_interrupts
120.71 ? 31% interrupts.CPU137.PMI:Performance_monitoring_interrupts
123.86 ? 33% interrupts.CPU138.NMI:Non-maskable_interrupts
123.86 ? 33% interrupts.CPU138.PMI:Performance_monitoring_interrupts
118.29 ? 44% interrupts.CPU139.NMI:Non-maskable_interrupts
118.29 ? 44% interrupts.CPU139.PMI:Performance_monitoring_interrupts
127.29 ? 73% interrupts.CPU142.NMI:Non-maskable_interrupts
127.29 ? 73% interrupts.CPU142.PMI:Performance_monitoring_interrupts
116.14 ? 29% interrupts.CPU147.NMI:Non-maskable_interrupts
116.14 ? 29% interrupts.CPU147.PMI:Performance_monitoring_interrupts
120.43 ? 32% interrupts.CPU148.NMI:Non-maskable_interrupts
120.43 ? 32% interrupts.CPU148.PMI:Performance_monitoring_interrupts
113.57 ? 31% interrupts.CPU149.NMI:Non-maskable_interrupts
113.57 ? 31% interrupts.CPU149.PMI:Performance_monitoring_interrupts
113.00 ? 32% interrupts.CPU162.NMI:Non-maskable_interrupts
113.00 ? 32% interrupts.CPU162.PMI:Performance_monitoring_interrupts
97.71 ? 35% interrupts.CPU172.NMI:Non-maskable_interrupts
97.71 ? 35% interrupts.CPU172.PMI:Performance_monitoring_interrupts
115.14 ? 39% interrupts.CPU174.NMI:Non-maskable_interrupts
115.14 ? 39% interrupts.CPU174.PMI:Performance_monitoring_interrupts
120.43 ? 32% interrupts.CPU178.NMI:Non-maskable_interrupts
120.43 ? 32% interrupts.CPU178.PMI:Performance_monitoring_interrupts
124.00 ? 35% interrupts.CPU180.NMI:Non-maskable_interrupts
124.00 ? 35% interrupts.CPU180.PMI:Performance_monitoring_interrupts
118.00 ? 31% interrupts.CPU182.NMI:Non-maskable_interrupts
118.00 ? 31% interrupts.CPU182.PMI:Performance_monitoring_interrupts
121.00 ? 33% interrupts.CPU183.NMI:Non-maskable_interrupts
121.00 ? 33% interrupts.CPU183.PMI:Performance_monitoring_interrupts
116.43 ? 30% interrupts.CPU184.NMI:Non-maskable_interrupts
116.43 ? 30% interrupts.CPU184.PMI:Performance_monitoring_interrupts
118.86 ? 31% interrupts.CPU185.NMI:Non-maskable_interrupts
118.86 ? 31% interrupts.CPU185.PMI:Performance_monitoring_interrupts
124.14 ? 33% interrupts.CPU29.NMI:Non-maskable_interrupts
124.14 ? 33% interrupts.CPU29.PMI:Performance_monitoring_interrupts
115.14 ? 40% interrupts.CPU30.NMI:Non-maskable_interrupts
115.14 ? 40% interrupts.CPU30.PMI:Performance_monitoring_interrupts
114.14 ? 40% interrupts.CPU31.NMI:Non-maskable_interrupts
114.14 ? 40% interrupts.CPU31.PMI:Performance_monitoring_interrupts
121.71 ? 32% interrupts.CPU33.NMI:Non-maskable_interrupts
121.71 ? 32% interrupts.CPU33.PMI:Performance_monitoring_interrupts
120.29 ? 33% interrupts.CPU34.NMI:Non-maskable_interrupts
120.29 ? 33% interrupts.CPU34.PMI:Performance_monitoring_interrupts
126.00 ? 35% interrupts.CPU35.NMI:Non-maskable_interrupts
126.00 ? 35% interrupts.CPU35.PMI:Performance_monitoring_interrupts
131.29 ? 38% interrupts.CPU36.NMI:Non-maskable_interrupts
131.29 ? 38% interrupts.CPU36.PMI:Performance_monitoring_interrupts
122.00 ? 33% interrupts.CPU37.NMI:Non-maskable_interrupts
122.00 ? 33% interrupts.CPU37.PMI:Performance_monitoring_interrupts
120.57 ? 33% interrupts.CPU38.NMI:Non-maskable_interrupts
120.57 ? 33% interrupts.CPU38.PMI:Performance_monitoring_interrupts
122.57 ? 32% interrupts.CPU39.NMI:Non-maskable_interrupts
122.57 ? 32% interrupts.CPU39.PMI:Performance_monitoring_interrupts
110.86 ? 35% interrupts.CPU41.NMI:Non-maskable_interrupts
110.86 ? 35% interrupts.CPU41.PMI:Performance_monitoring_interrupts
124.57 ? 33% interrupts.CPU42.NMI:Non-maskable_interrupts
124.57 ? 33% interrupts.CPU42.PMI:Performance_monitoring_interrupts
123.86 ? 34% interrupts.CPU43.NMI:Non-maskable_interrupts
123.86 ? 34% interrupts.CPU43.PMI:Performance_monitoring_interrupts
122.57 ? 33% interrupts.CPU44.NMI:Non-maskable_interrupts
122.57 ? 33% interrupts.CPU44.PMI:Performance_monitoring_interrupts
121.29 ? 32% interrupts.CPU45.NMI:Non-maskable_interrupts
121.29 ? 32% interrupts.CPU45.PMI:Performance_monitoring_interrupts
203.43 ?106% interrupts.CPU46.NMI:Non-maskable_interrupts
203.43 ?106% interrupts.CPU46.PMI:Performance_monitoring_interrupts
121.57 ? 33% interrupts.CPU47.NMI:Non-maskable_interrupts
121.57 ? 33% interrupts.CPU47.PMI:Performance_monitoring_interrupts
138.29 ? 29% interrupts.CPU50.NMI:Non-maskable_interrupts
138.29 ? 29% interrupts.CPU50.PMI:Performance_monitoring_interrupts
127.43 ? 27% interrupts.CPU51.NMI:Non-maskable_interrupts
127.43 ? 27% interrupts.CPU51.PMI:Performance_monitoring_interrupts
127.43 ? 27% interrupts.CPU52.NMI:Non-maskable_interrupts
127.43 ? 27% interrupts.CPU52.PMI:Performance_monitoring_interrupts
135.86 ? 37% interrupts.CPU54.NMI:Non-maskable_interrupts
135.86 ? 37% interrupts.CPU54.PMI:Performance_monitoring_interrupts
123.71 ? 30% interrupts.CPU55.NMI:Non-maskable_interrupts
123.71 ? 30% interrupts.CPU55.PMI:Performance_monitoring_interrupts
204.29 ? 86% interrupts.CPU56.NMI:Non-maskable_interrupts
204.29 ? 86% interrupts.CPU56.PMI:Performance_monitoring_interrupts
114.43 ? 35% interrupts.CPU59.NMI:Non-maskable_interrupts
114.43 ? 35% interrupts.CPU59.PMI:Performance_monitoring_interrupts
114.43 ? 34% interrupts.CPU60.NMI:Non-maskable_interrupts
114.43 ? 34% interrupts.CPU60.PMI:Performance_monitoring_interrupts
117.14 ? 40% interrupts.CPU61.NMI:Non-maskable_interrupts
117.14 ? 40% interrupts.CPU61.PMI:Performance_monitoring_interrupts
123.00 ? 31% interrupts.CPU62.NMI:Non-maskable_interrupts
123.00 ? 31% interrupts.CPU62.PMI:Performance_monitoring_interrupts
123.14 ? 30% interrupts.CPU63.NMI:Non-maskable_interrupts
123.14 ? 30% interrupts.CPU63.PMI:Performance_monitoring_interrupts
189.57 ? 81% interrupts.CPU64.NMI:Non-maskable_interrupts
189.57 ? 81% interrupts.CPU64.PMI:Performance_monitoring_interrupts
154.29 ? 55% interrupts.CPU67.NMI:Non-maskable_interrupts
154.29 ? 55% interrupts.CPU67.PMI:Performance_monitoring_interrupts
126.14 ? 28% interrupts.CPU68.NMI:Non-maskable_interrupts
126.14 ? 28% interrupts.CPU68.PMI:Performance_monitoring_interrupts
126.29 ? 28% interrupts.CPU70.NMI:Non-maskable_interrupts
126.29 ? 28% interrupts.CPU70.PMI:Performance_monitoring_interrupts
110.00 ? 35% interrupts.CPU77.NMI:Non-maskable_interrupts
110.00 ? 35% interrupts.CPU77.PMI:Performance_monitoring_interrupts
110.71 ? 35% interrupts.CPU78.NMI:Non-maskable_interrupts
110.71 ? 35% interrupts.CPU78.PMI:Performance_monitoring_interrupts
110.86 ? 33% interrupts.CPU79.NMI:Non-maskable_interrupts
110.86 ? 33% interrupts.CPU79.PMI:Performance_monitoring_interrupts
149.29 ? 46% interrupts.CPU80.NMI:Non-maskable_interrupts
149.29 ? 46% interrupts.CPU80.PMI:Performance_monitoring_interrupts
121.00 ? 31% interrupts.CPU81.NMI:Non-maskable_interrupts
121.00 ? 31% interrupts.CPU81.PMI:Performance_monitoring_interrupts
107.43 ? 34% interrupts.CPU92.NMI:Non-maskable_interrupts
107.43 ? 34% interrupts.CPU92.PMI:Performance_monitoring_interrupts
107.00 ? 34% interrupts.CPU93.NMI:Non-maskable_interrupts
107.00 ? 34% interrupts.CPU93.PMI:Performance_monitoring_interrupts
144.14 ? 76% interrupts.CPU94.NMI:Non-maskable_interrupts
144.14 ? 76% interrupts.CPU94.PMI:Performance_monitoring_interrupts
3114 ? 40% interrupts.CPU98.NMI:Non-maskable_interrupts
3114 ? 40% interrupts.CPU98.PMI:Performance_monitoring_interrupts
179728 ? 9% interrupts.NMI:Non-maskable_interrupts
179728 ? 9% interrupts.PMI:Performance_monitoring_interrupts
328.00 ? 86% interrupts.TLB:TLB_shootdowns

will-it-scale.16.processes

-------------------------------------------------+
.+.+.+.+.. .+.+.+.+ |
+ .+.+ +.+ |
+ .+.+. .+. |
+ +.+ |
|
|
|
|
|
|
O O O |
O O O O O O O O O O O O O O |
O O O O |
-------------------------------------------------+

will-it-scale.per_process_ops

-------------------------------------------------+
+.+.+.+.+.+..+.+.+.+.+.+ |
+ .. |
+.+.+.+.+.+ |
|
|
|
|
|
|
O O O |
O O O O O O O O O O O O O O O O |
O O |
O |
-------------------------------------------------+

will-it-scale.workload

-------------------------------------------------+
.+.+.+.+.. .+.+.+.+ |
+ .+.+ +.+ |
+ .+.+. .+. |
+ +.+ |
|
|
|
|
|
|
O O O |
O O O O O O O O O O O O O O |
O O O O |
-------------------------------------------------+

************************************************************************
threads 2 sockets Skylake with 192G memory
==============================================================
config/mode/nr_task/rootfs/tbox_group/test/testcase/ucode:
rhel-8.3/process/50%/debian-10.4-x86_64-20200603.cgz/lkp-skl-fpga01/futex3/will-it-scale/0x2006a0a
Remove the CONFIG_TRACE_IRQFLAGS")
4e737875727ee7a1a
-----------------
%stddev
\
-1.2% 1.038e+08 will-it-scale.52.processes
-1.2% 1995983 will-it-scale.per_process_ops
-1.2% 1.038e+08 will-it-scale.workload
32.04 ? 19% turbostat.PKG_%
1.21 ? 11% perf-profile.calltrace.cycles-pp.exit_to_user_mode_prepare.syscall_exit_to_user_mode.do_syscall_64.entry_SYSCALL_64_after_hwframe.syscall
1.22 ? 11% perf-profile.children.cycles-pp.exit_to_user_mode_prepare
0.09 ? 11% perf-profile.self.cycles-pp.rcu_nocb_flush_deferred_wakeup
1.08 ? 11% perf-profile.self.cycles-pp.exit_to_user_mode_prepare
-1.2% 7.418e+09 perf-stat.i.branch-instructions
-1.2% 1.089e+08 perf-stat.i.branch-misses
-1.2% 1.035e+08 perf-stat.i.dTLB-load-misses
-5.2% 1.048e+08 perf-stat.i.iTLB-load-misses
+4.4% 497.12 perf-stat.i.instructions-per-iTLB-miss
+4.6% 494.76 perf-stat.overall.instructions-per-iTLB-miss
-1.2% 7.394e+09 perf-stat.ps.branch-instructions
-1.3% 1.086e+08 perf-stat.ps.branch-misses
-1.2% 1.032e+08 perf-stat.ps.dTLB-load-misses
-5.2% 1.045e+08 perf-stat.ps.iTLB-load-misses
125.71 ? 50% interrupts.CPU102.RES:Rescheduling_interrupts
810.57 ? 41% interrupts.CPU18.CAL:Function_call_interrupts
124.00 ? 50% interrupts.CPU21.RES:Rescheduling_interrupts
203.29 ? 20% interrupts.CPU36.RES:Rescheduling_interrupts
129.43 ? 65% interrupts.CPU44.RES:Rescheduling_interrupts
672.29 ? 11% interrupts.CPU5.CAL:Function_call_interrupts
118.14 ? 66% interrupts.CPU57.RES:Rescheduling_interrupts
3917 ? 43% interrupts.CPU59.NMI:Non-maskable_interrupts
3917 ? 43% interrupts.CPU59.PMI:Performance_monitoring_interrupts
70.29 ? 50% interrupts.CPU63.RES:Rescheduling_interrupts
623.57 ? 15% interrupts.CPU73.CAL:Function_call_interrupts
202.29 ? 33% interrupts.CPU73.RES:Rescheduling_interrupts
16493 ? 20% softirqs.CPU102.RCU
19737 ? 20% softirqs.CPU21.RCU
27510 ? 30% softirqs.CPU21.SCHED
19600 ? 17% softirqs.CPU33.RCU
18512 ? 24% softirqs.CPU35.RCU
20065 ? 16% softirqs.CPU36.RCU
17662 ? 17% softirqs.CPU39.RCU
14858 ? 55% softirqs.CPU5.SCHED
15756 ? 50% softirqs.CPU50.SCHED
18618 ? 20% softirqs.CPU57.RCU
28576 ? 32% softirqs.CPU57.SCHED
15075 ? 6% softirqs.CPU63.RCU
17572 ? 47% softirqs.CPU73.SCHED
16895 ? 20% softirqs.CPU76.RCU
17988 ? 11% softirqs.CPU77.RCU
17407 ? 16% softirqs.CPU93.RCU
estimated based on internal Intel analysis and are provided
purposes only. Any difference in system hardware or software
may affect actual performance.
Open Source Technology Center
org/hyperkitty/list/lkp@lists.01.org">https://lists.01.org/hyperkitty/list/[email protected] Intel Corporation

Attachments:

(No filename) (36.59 kB)
config-5.13.0-rc3-00001-g21db66c4ff4e (176.78 kB)
job-script (8.07 kB)
job.yaml (5.40 kB)
reproduce (346.00 B)
Download all attachments

2021-05-28 05:00:56

by Steven Rostedt

[permalink] [raw]

Subject: Re: [irqflags] 21db66c4ff: will-it-scale.per_process_ops -4.5% regression

On Tue, 25 May 2021 15:03:54 +0800
kernel test robot <[email protected]> wrote:

> FYI, we noticed a -4.5% regression of will-it-scale.per_process_ops due to commit:

As Mark explained, your patch was not a nop. And the 0day bot just
showed you why we didn't do it the way you suggested.

-- Steve