| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <20191114025910.GA24535@xsang-OptiPlex-9020>
Date: Thu, 14 Nov 2019 10:59:10 +0800
From: kernel test robot <oliver.sang@...el.com>
To: Alexander Potapenko <glider@...gle.com>
Cc: Linus Torvalds <torvalds@...ux-foundation.org>,
Kees Cook <keescook@...omium.org>,
Christoph Lameter <cl@...ux.com>,
Masahiro Yamada <yamada.masahiro@...ionext.com>,
"Serge E. Hallyn" <serge@...lyn.com>,
Nick Desaulniers <ndesaulniers@...gle.com>,
Kostya Serebryany <kcc@...gle.com>,
Dmitry Vyukov <dvyukov@...gle.com>,
Sandeep Patil <sspatil@...roid.com>,
Laura Abbott <labbott@...hat.com>,
Randy Dunlap <rdunlap@...radead.org>,
Jann Horn <jannh@...gle.com>,
Mark Rutland <mark.rutland@....com>,
Marco Elver <elver@...gle.com>,
Andrew Morton <akpm@...ux-foundation.org>,
LKML <linux-kernel@...r.kernel.org>, lkp@...ts.01.org
Subject: [mm] 6471384af2: hackbench.throughput -4.9% regression
Greeting,
FYI, we noticed a -4.9% regression of hackbench.throughput due to commit:
commit: 6471384af2a6530696fc0203bafe4de41a23c9ef ("mm: security: introduce init_on_alloc=1 and init_on_free=1 boot options")
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git master
in testcase: hackbench
on test machine: 16 threads Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz with 32G memory
with following parameters:
nr_threads: 100%
mode: process
ipc: socket
cpufreq_governor: performance
ucode: 0xb8
test-description: Hackbench is both a benchmark and a stress test for the Linux kernel scheduler.
test-url: https://github.com/linux-test-project/ltp/blob/master/testcases/kernel/sched/cfs-scheduler/hackbench.c
In addition to that, the commit also has significant impact on the following tests:
+------------------+-------------------------------------------------------------------+
| testcase: change | hackbench: hackbench.throughput -4.3% regression |
| test machine | 16 threads Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz with 32G memory |
| test parameters | cpufreq_governor=performance |
| | mode=process |
| | nr_threads=100% |
| | ucode=0xb8 |
+------------------+-------------------------------------------------------------------+
If you fix the issue, kindly add following tag
Reported-by: kernel test robot <oliver.sang@...el.com>
Details are as below:
-------------------------------------------------------------------------------------------------->
To reproduce:
git clone https://github.com/intel/lkp-tests.git
cd lkp-tests
bin/lkp install job.yaml # job file is attached in this email
bin/lkp run job.yaml
=========================================================================================
compiler/cpufreq_governor/ipc/kconfig/mode/nr_threads/rootfs/tbox_group/testcase/ucode:
gcc-7/performance/socket/x86_64-rhel-7.2-clear_lck_7595/process/100%/clear-ota-25590-x86_64-2018-10-18.cgz/lkp-cfl-e1/hackbench/0xb8
commit:
ba5c5e4a5d ("arm64: move jump_label_init() before parse_early_param()")
6471384af2 ("mm: security: introduce init_on_alloc=1 and init_on_free=1 boot options")
ba5c5e4a5da443e8 6471384af2a6530696fc0203baf
---------------- ---------------------------
%stddev %change %stddev
\ | \
90709 -4.9% 86276 hackbench.throughput
108209 +42.0% 153632 hackbench.time.involuntary_context_switches
825777 -3.9% 793179 hackbench.time.minor_page_faults
712.02 -3.9% 684.38 hackbench.time.user_time
5.616e+08 -4.3% 5.376e+08 hackbench.workload
0.84 +0.1 0.95 ± 2% mpstat.cpu.-1.irq%
1347 ± 3% +12.7% 1518 ± 2% slabinfo.kmalloc-rcl-64.num_objs
3.49 +4.7% 3.65 turbostat.RAMWatt
396367 -1.3% 391171 vmstat.system.cs
14755342 -3.8% 14194624 proc-vmstat.numa_hit
14755342 -3.8% 14194624 proc-vmstat.numa_local
14815709 -3.7% 14261066 proc-vmstat.pgalloc_normal
14779482 -3.8% 14224922 proc-vmstat.pgfree
8627 ±172% +682.5% 67514 ± 65% sched_debug.cfs_rq:/.MIN_vruntime.avg
137821 ±172% +585.5% 944775 ± 59% sched_debug.cfs_rq:/.MIN_vruntime.max
33372 ±172% +620.8% 240555 ± 61% sched_debug.cfs_rq:/.MIN_vruntime.stddev
8627 ±172% +682.5% 67514 ± 65% sched_debug.cfs_rq:/.max_vruntime.avg
137821 ±172% +585.5% 944775 ± 59% sched_debug.cfs_rq:/.max_vruntime.max
33372 ±172% +620.8% 240555 ± 61% sched_debug.cfs_rq:/.max_vruntime.stddev
556.14 ± 4% -9.3% 504.21 ± 5% sched_debug.cpu.clock_task.stddev
1883 ± 36% -48.0% 979.75 ± 26% interrupts.133:IR-PCI-MSI.2097153-edge.eth1-TxRx-0
46373 ± 4% -10.1% 41692 ± 4% interrupts.CPU0.RES:Rescheduling_interrupts
1883 ± 36% -48.0% 979.75 ± 26% interrupts.CPU1.133:IR-PCI-MSI.2097153-edge.eth1-TxRx-0
45230 ± 5% -11.7% 39943 ± 4% interrupts.CPU1.RES:Rescheduling_interrupts
57716 ± 3% -12.3% 50601 interrupts.CPU12.RES:Rescheduling_interrupts
53566 ± 3% -11.5% 47399 interrupts.CPU13.RES:Rescheduling_interrupts
52007 ± 4% -14.0% 44710 ± 5% interrupts.CPU15.RES:Rescheduling_interrupts
46887 ± 3% -12.3% 41140 ± 6% interrupts.CPU2.RES:Rescheduling_interrupts
52269 ± 5% -12.1% 45958 ± 5% interrupts.CPU3.RES:Rescheduling_interrupts
57120 ± 7% -12.0% 50281 ± 4% interrupts.CPU4.RES:Rescheduling_interrupts
49253 ± 3% -11.5% 43586 ± 2% interrupts.CPU7.RES:Rescheduling_interrupts
820390 -10.2% 736882 interrupts.RES:Rescheduling_interrupts
44.93 -2.4% 43.85 perf-stat.i.MPKI
9.635e+09 -3.5% 9.3e+09 perf-stat.i.branch-instructions
1.106e+08 -4.2% 1.06e+08 perf-stat.i.branch-misses
13.44 +0.3 13.72 perf-stat.i.cache-miss-rate%
2.912e+08 -4.3% 2.786e+08 perf-stat.i.cache-misses
2.171e+09 -6.3% 2.034e+09 perf-stat.i.cache-references
397614 -1.4% 391925 perf-stat.i.context-switches
1.26 +4.2% 1.31 perf-stat.i.cpi
62650 -3.9% 60222 perf-stat.i.cpu-migrations
209.19 +4.6% 218.81 perf-stat.i.cycles-between-cache-misses
1.459e+10 -4.0% 1.401e+10 perf-stat.i.dTLB-loads
1.025e+10 -4.2% 9.817e+09 perf-stat.i.dTLB-stores
86435026 -2.5% 84314445 perf-stat.i.iTLB-load-misses
4.833e+10 -4.0% 4.639e+10 perf-stat.i.instructions
562.26 -1.5% 553.60 perf-stat.i.instructions-per-iTLB-miss
0.79 -4.0% 0.76 perf-stat.i.ipc
0.00 ± 35% +0.0 0.00 ± 49% perf-stat.i.node-load-miss-rate%
0.64 ± 41% +40.2% 0.90 ± 32% perf-stat.i.node-load-misses
21652012 -2.2% 21184023 perf-stat.i.node-loads
0.57 ± 7% +76.9% 1.00 ± 33% perf-stat.i.node-store-misses
26105714 +79.9% 46970933 perf-stat.i.node-stores
44.93 -2.4% 43.85 perf-stat.overall.MPKI
13.41 +0.3 13.69 perf-stat.overall.cache-miss-rate%
1.26 +4.2% 1.31 perf-stat.overall.cpi
208.98 +4.6% 218.59 perf-stat.overall.cycles-between-cache-misses
559.11 -1.6% 550.27 perf-stat.overall.instructions-per-iTLB-miss
0.79 -4.0% 0.76 perf-stat.overall.ipc
0.00 ± 42% +0.0 0.00 ± 33% perf-stat.overall.node-load-miss-rate%
9.619e+09 -3.5% 9.285e+09 perf-stat.ps.branch-instructions
1.104e+08 -4.2% 1.058e+08 perf-stat.ps.branch-misses
2.907e+08 -4.3% 2.781e+08 perf-stat.ps.cache-misses
2.168e+09 -6.3% 2.031e+09 perf-stat.ps.cache-references
396957 -1.4% 391281 perf-stat.ps.context-switches
62546 -3.9% 60123 perf-stat.ps.cpu-migrations
1.457e+10 -4.0% 1.399e+10 perf-stat.ps.dTLB-loads
1.023e+10 -4.2% 9.801e+09 perf-stat.ps.dTLB-stores
86292583 -2.5% 84176193 perf-stat.ps.iTLB-load-misses
4.825e+10 -4.0% 4.632e+10 perf-stat.ps.instructions
0.64 ± 41% +40.2% 0.90 ± 32% perf-stat.ps.node-load-misses
21616330 -2.2% 21149342 perf-stat.ps.node-loads
0.56 ± 7% +76.9% 1.00 ± 33% perf-stat.ps.node-store-misses
26062702 +79.9% 46894016 perf-stat.ps.node-stores
2.929e+13 -3.5% 2.827e+13 perf-stat.total.instructions
39.46 -1.3 38.13 perf-profile.calltrace.cycles-pp.__x64_sys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe
39.37 -1.3 38.05 perf-profile.calltrace.cycles-pp.ksys_read.__x64_sys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe
38.18 -1.3 36.91 perf-profile.calltrace.cycles-pp.vfs_read.ksys_read.__x64_sys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe
35.37 -1.2 34.19 perf-profile.calltrace.cycles-pp.__vfs_read.vfs_read.ksys_read.__x64_sys_read.do_syscall_64
34.98 -1.2 33.81 perf-profile.calltrace.cycles-pp.new_sync_read.__vfs_read.vfs_read.ksys_read.__x64_sys_read
34.12 -1.1 33.00 perf-profile.calltrace.cycles-pp.sock_read_iter.new_sync_read.__vfs_read.vfs_read.ksys_read
32.88 -1.0 31.84 perf-profile.calltrace.cycles-pp.sock_recvmsg.sock_read_iter.new_sync_read.__vfs_read.vfs_read
32.46 -1.0 31.43 perf-profile.calltrace.cycles-pp.unix_stream_recvmsg.sock_recvmsg.sock_read_iter.new_sync_read.__vfs_read
32.04 -1.0 31.03 perf-profile.calltrace.cycles-pp.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg.sock_read_iter.new_sync_read
2.70 ± 2% -0.8 1.93 perf-profile.calltrace.cycles-pp._raw_spin_lock.unix_stream_sendmsg.sock_sendmsg.sock_write_iter.new_sync_write
2.16 ± 2% -0.8 1.41 perf-profile.calltrace.cycles-pp.refcount_add_checked.skb_set_owner_w.sock_alloc_send_pskb.unix_stream_sendmsg.sock_sendmsg
2.70 -0.7 1.95 perf-profile.calltrace.cycles-pp.skb_set_owner_w.sock_alloc_send_pskb.unix_stream_sendmsg.sock_sendmsg.sock_write_iter
2.16 ± 2% -0.7 1.41 perf-profile.calltrace.cycles-pp.refcount_add_not_zero_checked.refcount_add_checked.skb_set_owner_w.sock_alloc_send_pskb.unix_stream_sendmsg
11.94 -0.4 11.50 perf-profile.calltrace.cycles-pp.consume_skb.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg.sock_read_iter
9.59 -0.4 9.18 perf-profile.calltrace.cycles-pp.unix_stream_read_actor.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg.sock_read_iter
9.49 -0.4 9.09 perf-profile.calltrace.cycles-pp.skb_copy_datagram_iter.unix_stream_read_actor.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg
9.20 -0.4 8.85 perf-profile.calltrace.cycles-pp.__skb_datagram_iter.skb_copy_datagram_iter.unix_stream_read_actor.unix_stream_read_generic.unix_stream_recvmsg
8.85 -0.3 8.54 perf-profile.calltrace.cycles-pp.simple_copy_to_iter.__skb_datagram_iter.skb_copy_datagram_iter.unix_stream_read_actor.unix_stream_read_generic
6.51 -0.3 6.21 perf-profile.calltrace.cycles-pp._copy_to_iter.simple_copy_to_iter.__skb_datagram_iter.skb_copy_datagram_iter.unix_stream_read_actor
7.29 -0.3 7.00 perf-profile.calltrace.cycles-pp.skb_release_all.consume_skb.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg
5.44 -0.2 5.20 perf-profile.calltrace.cycles-pp.copy_user_enhanced_fast_string._copy_to_iter.simple_copy_to_iter.__skb_datagram_iter.skb_copy_datagram_iter
3.99 -0.2 3.79 perf-profile.calltrace.cycles-pp.syscall_slow_exit_work.do_syscall_64.entry_SYSCALL_64_after_hwframe
2.44 ± 2% -0.2 2.24 ± 4% perf-profile.calltrace.cycles-pp.___cache_free.kfree.skb_free_head.skb_release_data.skb_release_all
3.23 -0.2 3.05 perf-profile.calltrace.cycles-pp.skb_release_head_state.skb_release_all.consume_skb.unix_stream_read_generic.unix_stream_recvmsg
3.28 -0.2 3.10 perf-profile.calltrace.cycles-pp.__audit_syscall_exit.syscall_slow_exit_work.do_syscall_64.entry_SYSCALL_64_after_hwframe
2.74 -0.2 2.56 perf-profile.calltrace.cycles-pp.skb_queue_tail.unix_stream_sendmsg.sock_sendmsg.sock_write_iter.new_sync_write
3.05 -0.2 2.88 perf-profile.calltrace.cycles-pp.unix_destruct_scm.skb_release_head_state.skb_release_all.consume_skb.unix_stream_read_generic
1.53 ± 3% -0.2 1.36 perf-profile.calltrace.cycles-pp.syscall_return_via_sysret
4.44 -0.2 4.27 perf-profile.calltrace.cycles-pp.skb_copy_datagram_from_iter.unix_stream_sendmsg.sock_sendmsg.sock_write_iter.new_sync_write
2.62 -0.2 2.46 perf-profile.calltrace.cycles-pp.sock_wfree.unix_destruct_scm.skb_release_head_state.skb_release_all.consume_skb
1.95 -0.2 1.79 perf-profile.calltrace.cycles-pp._raw_spin_lock_irqsave.skb_queue_tail.unix_stream_sendmsg.sock_sendmsg.sock_write_iter
1.11 -0.1 0.97 ± 3% perf-profile.calltrace.cycles-pp._raw_spin_lock.___cache_free.kfree.skb_free_head.skb_release_data
1.53 ± 2% -0.1 1.39 ± 2% perf-profile.calltrace.cycles-pp.cache_alloc_refill.kmem_cache_alloc_node_trace.__kmalloc_node_track_caller.__kmalloc_reserve.__alloc_skb
1.02 ± 2% -0.1 0.89 ± 3% perf-profile.calltrace.cycles-pp.queued_spin_lock_slowpath._raw_spin_lock.___cache_free.kfree.skb_free_head
3.69 -0.1 3.56 perf-profile.calltrace.cycles-pp.kmem_cache_free.kfree_skbmem.consume_skb.unix_stream_read_generic.unix_stream_recvmsg
3.83 -0.1 3.70 perf-profile.calltrace.cycles-pp.kfree_skbmem.consume_skb.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg
1.60 -0.1 1.47 ± 2% perf-profile.calltrace.cycles-pp.entry_SYSCALL_64
1.85 -0.1 1.73 perf-profile.calltrace.cycles-pp.syscall_trace_enter.do_syscall_64.entry_SYSCALL_64_after_hwframe
1.60 -0.1 1.48 ± 2% perf-profile.calltrace.cycles-pp.skb_unlink.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg.sock_read_iter
2.70 -0.1 2.59 perf-profile.calltrace.cycles-pp._copy_from_iter.skb_copy_datagram_from_iter.unix_stream_sendmsg.sock_sendmsg.sock_write_iter
0.69 ± 3% -0.1 0.58 perf-profile.calltrace.cycles-pp.queued_spin_lock_slowpath._raw_spin_lock.cache_alloc_refill.kmem_cache_alloc_node_trace.__kmalloc_node_track_caller
1.77 -0.1 1.66 perf-profile.calltrace.cycles-pp.___cache_free.kmem_cache_free.kfree_skbmem.consume_skb.unix_stream_read_generic
0.74 ± 2% -0.1 0.64 perf-profile.calltrace.cycles-pp._raw_spin_lock.cache_alloc_refill.kmem_cache_alloc_node_trace.__kmalloc_node_track_caller.__kmalloc_reserve
1.82 -0.1 1.72 perf-profile.calltrace.cycles-pp.__fget_light.__fdget_pos.ksys_write.__x64_sys_write.do_syscall_64
1.90 -0.1 1.81 perf-profile.calltrace.cycles-pp.__fdget_pos.ksys_write.__x64_sys_write.do_syscall_64.entry_SYSCALL_64_after_hwframe
0.98 ± 2% -0.1 0.89 ± 3% perf-profile.calltrace.cycles-pp.unroll_tree_refs.__audit_syscall_exit.syscall_slow_exit_work.do_syscall_64.entry_SYSCALL_64_after_hwframe
1.05 -0.1 0.99 ± 3% perf-profile.calltrace.cycles-pp.unix_write_space.sock_wfree.unix_destruct_scm.skb_release_head_state.skb_release_all
0.97 ± 2% -0.1 0.92 ± 2% perf-profile.calltrace.cycles-pp.__audit_syscall_entry.syscall_trace_enter.do_syscall_64.entry_SYSCALL_64_after_hwframe
0.65 -0.1 0.59 ± 3% perf-profile.calltrace.cycles-pp._raw_spin_lock.___cache_free.kmem_cache_free.kfree_skbmem.consume_skb
0.60 -0.1 0.55 ± 3% perf-profile.calltrace.cycles-pp.queued_spin_lock_slowpath._raw_spin_lock.___cache_free.kmem_cache_free.kfree_skbmem
0.70 -0.1 0.64 perf-profile.calltrace.cycles-pp._raw_spin_lock_irqsave.skb_unlink.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg
0.86 -0.1 0.81 perf-profile.calltrace.cycles-pp.free_block.___cache_free.kmem_cache_free.kfree_skbmem.consume_skb
2.03 -0.1 1.98 perf-profile.calltrace.cycles-pp.schedule_timeout.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg.sock_read_iter
0.66 ± 2% -0.1 0.60 perf-profile.calltrace.cycles-pp.refcount_inc_not_zero_checked.refcount_inc_checked.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg
1.98 -0.0 1.94 perf-profile.calltrace.cycles-pp.schedule.schedule_timeout.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg
0.69 ± 3% -0.0 0.64 ± 2% perf-profile.calltrace.cycles-pp.refcount_inc_checked.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg.sock_read_iter
1.94 -0.0 1.89 perf-profile.calltrace.cycles-pp.__sched_text_start.schedule.schedule_timeout.unix_stream_read_generic.unix_stream_recvmsg
0.75 ± 3% -0.0 0.71 perf-profile.calltrace.cycles-pp.deactivate_task.__sched_text_start.schedule.schedule_timeout.unix_stream_read_generic
0.70 ± 2% -0.0 0.67 perf-profile.calltrace.cycles-pp.cache_alloc_refill.kmem_cache_alloc_node.__alloc_skb.alloc_skb_with_frags.sock_alloc_send_pskb
1.86 ± 2% +0.2 2.08 perf-profile.calltrace.cycles-pp.refcount_inc_not_zero_checked.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg.sock_read_iter
92.65 +0.4 93.01 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe
90.21 +0.6 90.79 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe
0.00 +1.0 1.00 ± 4% perf-profile.calltrace.cycles-pp.memset_erms.__alloc_skb.alloc_skb_with_frags.sock_alloc_send_pskb.unix_stream_sendmsg
0.00 +1.1 1.06 perf-profile.calltrace.cycles-pp.memset_erms.__kmalloc_node_track_caller.__kmalloc_reserve.__alloc_skb.alloc_skb_with_frags
4.41 +1.7 6.13 perf-profile.calltrace.cycles-pp.kmem_cache_alloc_node_trace.__kmalloc_node_track_caller.__kmalloc_reserve.__alloc_skb.alloc_skb_with_frags
42.73 +2.2 44.95 perf-profile.calltrace.cycles-pp.__x64_sys_write.do_syscall_64.entry_SYSCALL_64_after_hwframe
42.64 +2.2 44.86 perf-profile.calltrace.cycles-pp.ksys_write.__x64_sys_write.do_syscall_64.entry_SYSCALL_64_after_hwframe
40.20 +2.4 42.55 perf-profile.calltrace.cycles-pp.vfs_write.ksys_write.__x64_sys_write.do_syscall_64.entry_SYSCALL_64_after_hwframe
37.99 +2.4 40.41 perf-profile.calltrace.cycles-pp.__vfs_write.vfs_write.ksys_write.__x64_sys_write.do_syscall_64
37.73 +2.4 40.18 perf-profile.calltrace.cycles-pp.new_sync_write.__vfs_write.vfs_write.ksys_write.__x64_sys_write
36.93 +2.5 39.41 perf-profile.calltrace.cycles-pp.sock_write_iter.new_sync_write.__vfs_write.vfs_write.ksys_write
35.68 +2.5 38.19 perf-profile.calltrace.cycles-pp.sock_sendmsg.sock_write_iter.new_sync_write.__vfs_write.vfs_write
34.85 +2.5 37.36 perf-profile.calltrace.cycles-pp.unix_stream_sendmsg.sock_sendmsg.sock_write_iter.new_sync_write.__vfs_write
4.99 +2.8 7.79 perf-profile.calltrace.cycles-pp.__kmalloc_node_track_caller.__kmalloc_reserve.__alloc_skb.alloc_skb_with_frags.sock_alloc_send_pskb
5.22 +3.0 8.21 perf-profile.calltrace.cycles-pp.__kmalloc_reserve.__alloc_skb.alloc_skb_with_frags.sock_alloc_send_pskb.unix_stream_sendmsg
16.62 +4.0 20.60 perf-profile.calltrace.cycles-pp.sock_alloc_send_pskb.unix_stream_sendmsg.sock_sendmsg.sock_write_iter.new_sync_write
12.22 +4.7 16.95 perf-profile.calltrace.cycles-pp.__alloc_skb.alloc_skb_with_frags.sock_alloc_send_pskb.unix_stream_sendmsg.sock_sendmsg
12.55 +4.7 17.29 perf-profile.calltrace.cycles-pp.alloc_skb_with_frags.sock_alloc_send_pskb.unix_stream_sendmsg.sock_sendmsg.sock_write_iter
39.76 -1.3 38.41 perf-profile.children.cycles-pp.__x64_sys_read
39.48 -1.3 38.16 perf-profile.children.cycles-pp.ksys_read
38.25 -1.3 36.99 perf-profile.children.cycles-pp.vfs_read
35.48 -1.2 34.28 perf-profile.children.cycles-pp.__vfs_read
35.06 -1.2 33.89 perf-profile.children.cycles-pp.new_sync_read
34.37 -1.1 33.23 perf-profile.children.cycles-pp.sock_read_iter
6.75 -1.1 5.62 perf-profile.children.cycles-pp._raw_spin_lock
32.74 -1.0 31.70 perf-profile.children.cycles-pp.unix_stream_recvmsg
33.00 -1.0 31.95 perf-profile.children.cycles-pp.sock_recvmsg
32.16 -1.0 31.14 perf-profile.children.cycles-pp.unix_stream_read_generic
2.74 -0.8 1.98 perf-profile.children.cycles-pp.skb_set_owner_w
2.21 ± 2% -0.7 1.46 perf-profile.children.cycles-pp.refcount_add_checked
2.33 -0.7 1.59 perf-profile.children.cycles-pp.refcount_add_not_zero_checked
12.06 -0.4 11.61 perf-profile.children.cycles-pp.consume_skb
9.63 -0.4 9.22 perf-profile.children.cycles-pp.unix_stream_read_actor
9.56 -0.4 9.15 perf-profile.children.cycles-pp.skb_copy_datagram_iter
9.26 -0.4 8.90 perf-profile.children.cycles-pp.__skb_datagram_iter
8.99 -0.4 8.63 perf-profile.children.cycles-pp.simple_copy_to_iter
4.35 -0.4 3.99 perf-profile.children.cycles-pp.queued_spin_lock_slowpath
7.37 -0.3 7.06 perf-profile.children.cycles-pp.skb_release_all
4.26 ± 2% -0.3 3.95 perf-profile.children.cycles-pp.___cache_free
6.56 -0.3 6.26 perf-profile.children.cycles-pp._copy_to_iter
7.13 -0.3 6.83 perf-profile.children.cycles-pp.copy_user_enhanced_fast_string
4.06 -0.3 3.79 perf-profile.children.cycles-pp._raw_spin_lock_irqsave
3.06 -0.3 2.80 perf-profile.children.cycles-pp.syscall_return_via_sysret
4.12 -0.2 3.91 perf-profile.children.cycles-pp.syscall_slow_exit_work
2.78 -0.2 2.59 perf-profile.children.cycles-pp.skb_queue_tail
2.46 ± 2% -0.2 2.27 perf-profile.children.cycles-pp.__might_sleep
3.28 -0.2 3.10 perf-profile.children.cycles-pp.skb_release_head_state
2.74 -0.2 2.57 perf-profile.children.cycles-pp.sock_wfree
3.31 -0.2 3.14 ± 2% perf-profile.children.cycles-pp.unix_destruct_scm
4.49 -0.2 4.32 perf-profile.children.cycles-pp.skb_copy_datagram_from_iter
3.48 -0.2 3.31 perf-profile.children.cycles-pp.__audit_syscall_exit
2.24 -0.2 2.08 perf-profile.children.cycles-pp.cache_alloc_refill
3.79 -0.2 3.64 perf-profile.children.cycles-pp.kmem_cache_free
3.90 -0.1 3.77 perf-profile.children.cycles-pp.kfree_skbmem
1.62 -0.1 1.49 perf-profile.children.cycles-pp.entry_SYSCALL_64
2.77 -0.1 2.65 perf-profile.children.cycles-pp._copy_from_iter
1.94 -0.1 1.82 perf-profile.children.cycles-pp.syscall_trace_enter
2.52 -0.1 2.40 perf-profile.children.cycles-pp.__fget_light
3.14 ± 2% -0.1 3.02 perf-profile.children.cycles-pp.__check_object_size
1.58 ± 3% -0.1 1.46 perf-profile.children.cycles-pp.___might_sleep
1.61 -0.1 1.50 ± 2% perf-profile.children.cycles-pp.skb_unlink
2.67 -0.1 2.56 perf-profile.children.cycles-pp.__fdget_pos
1.77 ± 2% -0.1 1.67 perf-profile.children.cycles-pp.free_block
1.10 ± 3% -0.1 1.01 ± 3% perf-profile.children.cycles-pp.__audit_syscall_entry
0.57 ± 5% -0.1 0.48 ± 9% perf-profile.children.cycles-pp.memcg_kmem_put_cache
1.08 ± 2% -0.1 1.00 ± 2% perf-profile.children.cycles-pp.unroll_tree_refs
0.93 ± 4% -0.1 0.84 perf-profile.children.cycles-pp.__might_fault
0.18 ± 7% -0.1 0.11 ± 4% perf-profile.children.cycles-pp.should_failslab
0.56 ± 6% -0.1 0.50 perf-profile.children.cycles-pp.rcu_all_qs
2.70 -0.1 2.64 perf-profile.children.cycles-pp.schedule
2.75 -0.1 2.69 perf-profile.children.cycles-pp.schedule_timeout
2.83 -0.1 2.78 perf-profile.children.cycles-pp.__sched_text_start
1.07 -0.1 1.02 perf-profile.children.cycles-pp.__list_del_entry_valid
0.70 ± 3% -0.1 0.65 ± 2% perf-profile.children.cycles-pp.refcount_inc_checked
0.30 -0.0 0.25 ± 6% perf-profile.children.cycles-pp.copyin
0.67 -0.0 0.62 perf-profile.children.cycles-pp._cond_resched
1.19 -0.0 1.15 ± 2% perf-profile.children.cycles-pp.security_file_permission
0.74 ± 2% -0.0 0.70 ± 2% perf-profile.children.cycles-pp.wait_for_unix_gc
1.14 -0.0 1.10 ± 2% perf-profile.children.cycles-pp.unix_write_space
0.32 ± 2% -0.0 0.29 ± 3% perf-profile.children.cycles-pp.__x86_indirect_thunk_rax
0.98 -0.0 0.94 perf-profile.children.cycles-pp.deactivate_task
0.22 -0.0 0.20 ± 5% perf-profile.children.cycles-pp.update_rq_clock
0.73 -0.0 0.71 ± 2% perf-profile.children.cycles-pp.dequeue_task_fair
0.16 ± 2% -0.0 0.15 perf-profile.children.cycles-pp.set_next_entity
0.08 ± 8% +0.0 0.10 ± 4% perf-profile.children.cycles-pp.cache_grow_begin
0.09 ± 4% +0.0 0.12 ± 11% perf-profile.children.cycles-pp.maybe_add_creds
0.00 +0.1 0.08 ± 6% perf-profile.children.cycles-pp.memset
0.97 ± 5% +0.2 1.13 ± 3% perf-profile.children.cycles-pp.__virt_addr_valid
2.52 ± 2% +0.2 2.69 perf-profile.children.cycles-pp.refcount_inc_not_zero_checked
92.69 +0.4 93.05 perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe
90.39 +0.6 90.97 perf-profile.children.cycles-pp.do_syscall_64
4.64 +1.8 6.49 perf-profile.children.cycles-pp.kmem_cache_alloc_node_trace
0.00 +2.1 2.11 ± 2% perf-profile.children.cycles-pp.memset_erms
43.03 +2.2 45.23 perf-profile.children.cycles-pp.__x64_sys_write
42.76 +2.2 44.96 perf-profile.children.cycles-pp.ksys_write
40.25 +2.3 42.59 perf-profile.children.cycles-pp.vfs_write
38.06 +2.4 40.48 perf-profile.children.cycles-pp.__vfs_write
37.80 +2.4 40.24 perf-profile.children.cycles-pp.new_sync_write
37.11 +2.5 39.57 perf-profile.children.cycles-pp.sock_write_iter
35.75 +2.5 38.24 perf-profile.children.cycles-pp.sock_sendmsg
35.18 +2.5 37.70 perf-profile.children.cycles-pp.unix_stream_sendmsg
5.05 +2.8 7.89 perf-profile.children.cycles-pp.__kmalloc_node_track_caller
5.26 +3.0 8.30 perf-profile.children.cycles-pp.__kmalloc_reserve
16.66 +4.0 20.62 perf-profile.children.cycles-pp.sock_alloc_send_pskb
12.30 +4.7 17.02 perf-profile.children.cycles-pp.__alloc_skb
12.59 +4.7 17.33 perf-profile.children.cycles-pp.alloc_skb_with_frags
3.43 ± 2% -0.8 2.61 perf-profile.self.cycles-pp._raw_spin_lock
2.31 -0.7 1.57 perf-profile.self.cycles-pp.refcount_add_not_zero_checked
4.33 -0.3 3.98 perf-profile.self.cycles-pp.queued_spin_lock_slowpath
7.10 -0.3 6.80 perf-profile.self.cycles-pp.copy_user_enhanced_fast_string
3.03 -0.3 2.77 perf-profile.self.cycles-pp.syscall_return_via_sysret
1.87 -0.2 1.62 ± 2% perf-profile.self.cycles-pp.__check_object_size
3.01 -0.2 2.80 perf-profile.self.cycles-pp._raw_spin_lock_irqsave
1.76 -0.2 1.59 perf-profile.self.cycles-pp.entry_SYSCALL_64_after_hwframe
1.37 ± 2% -0.1 1.24 ± 2% perf-profile.self.cycles-pp.sock_def_readable
1.62 -0.1 1.49 perf-profile.self.cycles-pp.entry_SYSCALL_64
2.46 -0.1 2.34 perf-profile.self.cycles-pp.__fget_light
1.80 -0.1 1.69 perf-profile.self.cycles-pp.unix_stream_read_generic
1.49 ± 2% -0.1 1.38 ± 2% perf-profile.self.cycles-pp.___might_sleep
0.90 ± 3% -0.1 0.81 ± 2% perf-profile.self.cycles-pp.__audit_syscall_entry
1.07 ± 2% -0.1 0.98 ± 3% perf-profile.self.cycles-pp.unroll_tree_refs
1.14 ± 2% -0.1 1.06 perf-profile.self.cycles-pp.sock_read_iter
0.66 ± 3% -0.1 0.58 ± 2% perf-profile.self.cycles-pp.sock_wfree
1.04 ± 3% -0.1 0.98 perf-profile.self.cycles-pp.__might_sleep
1.04 -0.1 0.98 perf-profile.self.cycles-pp.__list_del_entry_valid
1.86 -0.1 1.80 perf-profile.self.cycles-pp.__audit_syscall_exit
0.46 ± 8% -0.1 0.40 ± 3% perf-profile.self.cycles-pp.rcu_all_qs
0.98 ± 2% -0.1 0.93 ± 4% perf-profile.self.cycles-pp.free_block
0.74 ± 2% -0.1 0.69 ± 2% perf-profile.self.cycles-pp.vfs_read
0.58 -0.0 0.54 ± 2% perf-profile.self.cycles-pp.syscall_slow_exit_work
0.11 ± 9% -0.0 0.07 ± 5% perf-profile.self.cycles-pp.should_failslab
0.74 -0.0 0.70 perf-profile.self.cycles-pp.new_sync_read
0.37 ± 3% -0.0 0.34 ± 2% perf-profile.self.cycles-pp._cond_resched
0.64 -0.0 0.60 perf-profile.self.cycles-pp._copy_from_iter
0.18 ± 2% -0.0 0.15 ± 5% perf-profile.self.cycles-pp.copyin
0.30 ± 5% -0.0 0.27 ± 5% perf-profile.self.cycles-pp.__skb_datagram_iter
0.25 ± 4% -0.0 0.23 ± 6% perf-profile.self.cycles-pp.simple_copy_to_iter
0.23 -0.0 0.21 ± 5% perf-profile.self.cycles-pp.sock_sendmsg
0.14 ± 5% -0.0 0.12 perf-profile.self.cycles-pp.__list_add_valid
0.18 -0.0 0.16 ± 4% perf-profile.self.cycles-pp.__x86_indirect_thunk_rax
0.12 ± 4% -0.0 0.11 ± 6% perf-profile.self.cycles-pp.security_socket_sendmsg
0.09 ± 4% +0.0 0.11 ± 9% perf-profile.self.cycles-pp.maybe_add_creds
0.11 ± 21% +0.0 0.15 ± 24% perf-profile.self.cycles-pp.kmalloc_slab
0.19 ± 5% +0.1 0.26 ± 12% perf-profile.self.cycles-pp.__kmalloc_reserve
1.11 ± 2% +0.1 1.19 perf-profile.self.cycles-pp.kfree
0.13 ± 11% +0.1 0.22 ± 8% perf-profile.self.cycles-pp.__kmalloc_node_track_caller
0.94 ± 5% +0.1 1.09 ± 3% perf-profile.self.cycles-pp.__virt_addr_valid
2.51 ± 2% +0.2 2.68 perf-profile.self.cycles-pp.refcount_inc_not_zero_checked
1.66 ± 5% +0.8 2.42 perf-profile.self.cycles-pp.__alloc_skb
1.26 +1.9 3.15 ± 2% perf-profile.self.cycles-pp.kmem_cache_alloc_node_trace
0.00 +2.1 2.07 ± 2% perf-profile.self.cycles-pp.memset_erms
hackbench.throughput
100000 +-+----------------------------------------------------------------+
90000 +-+++ +.++ +.++.++.+ + ++.+ ++ +.++ +.++ +.+ ++ +.+ + +|
|O OO OO O: O O O O OO :O O OO O O: OO O O: OO :O O: O OO :|
80000 +-+ : : : : : : : : :: : : : : : : :: : : : :|
70000 +-+ : : : : : : : : :: : : : : : : :: : : : :|
|: : : : : : :: : : : : : : : : : : : : : : :: : |
60000 +-+ : : : : : :: : : : : : : : : : : : : : : :: : |
50000 +-+ : : : : : :: : : : : : : : : : : : : : : :: : |
40000 +-+ :: :: :: :: :: :: :: :: :: :: :: :: |
| :: :: :: :: :: :: :: :: :: :: :: :: |
30000 +-+ :: :: :: :: :: :: :: :: :: :: :: :: |
20000 +-+ : : : : : : : : : : : : |
| : : : : : : : : : : : : |
10000 +-+ : : : : : : : : : : : : |
0 O-+-------O-O---O-O---O----O---O----O---O----O---O----O---O--O-----+
hackbench.time.involuntary_context_switches
180000 +-+----------------------------------------------------------------+
| O O O OO O |
160000 +O+ O O O O O O O OO O O O O O O O OO |
140000 +-+O O OO |
| + |
120000 +-+ + +.++.++.+ + : +. + |
100000 +-+++ +.+: : : : ++.+ ++ +.++ +.+: : + :+ +.+ + +|
|: : : : : : : : : :: : : : : : : :: : : : :|
80000 +-+ : : : : : :: : : : : : : : : : : : : : : : :|
60000 +-+ : : : : : :: : : : : : : : : : : : : : : :: : |
|: : : :: :: :: :: : : : : :: :: : : :: :: |
40000 +-+ :: :: :: :: :: :: :: :: :: :: :: :: |
20000 +-+ : : : :: :: : : : : : :: :: |
| : : : : : : : : : : : : |
0 O-+-------O-O---O-O---O----O---O----O---O----O---O----O---O--O-----+
[*] bisect-good sample
[O] bisect-bad sample
***************************************************************************************************
lkp-cfl-e1: 16 threads Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz with 32G memory
=========================================================================================
compiler/cpufreq_governor/kconfig/mode/nr_threads/rootfs/tbox_group/testcase/ucode:
gcc-7/performance/x86_64-rhel-7.2-clear_lck_7595/process/100%/clear-ota-25590-x86_64-2018-10-18.cgz/lkp-cfl-e1/hackbench/0xb8
commit:
ba5c5e4a5d ("arm64: move jump_label_init() before parse_early_param()")
6471384af2 ("mm: security: introduce init_on_alloc=1 and init_on_free=1 boot options")
ba5c5e4a5da443e8 6471384af2a6530696fc0203baf
---------------- ---------------------------
%stddev %change %stddev
\ | \
88494 -4.3% 84657 hackbench.throughput
138438 ± 4% +31.9% 182656 ± 7% hackbench.time.involuntary_context_switches
694.79 -5.9% 653.85 hackbench.time.user_time
5.568e+08 -4.3% 5.328e+08 hackbench.workload
8544005 ± 38% -45.2% 4685694 ± 6% cpuidle.C1E.time
5.31 ± 11% +29.2% 6.86 ± 16% sched_debug.cpu.nr_uninterruptible.stddev
3.48 +4.6% 3.64 turbostat.RAMWatt
0.84 ± 2% +0.1 0.95 mpstat.cpu.-1.irq%
0.05 ± 3% +0.0 0.06 ± 10% mpstat.cpu.-1.soft%
15632269 -4.1% 14984464 ± 3% proc-vmstat.numa_hit
15632269 -4.1% 14984464 ± 3% proc-vmstat.numa_local
15715101 -4.2% 15056704 ± 3% proc-vmstat.pgalloc_normal
15686075 -4.2% 15022294 ± 3% proc-vmstat.pgfree
45.36 -2.9% 44.03 perf-stat.i.MPKI
9.663e+09 -2.7% 9.402e+09 perf-stat.i.branch-instructions
1.088e+08 -2.7% 1.058e+08 perf-stat.i.branch-misses
13.34 +0.3 13.62 perf-stat.i.cache-miss-rate%
2.908e+08 -4.1% 2.787e+08 perf-stat.i.cache-misses
2.179e+09 -6.0% 2.048e+09 perf-stat.i.cache-references
1.26 +3.4% 1.31 perf-stat.i.cpi
61720 -2.9% 59924 perf-stat.i.cpu-migrations
210.15 +4.0% 218.46 perf-stat.i.cycles-between-cache-misses
1.449e+10 -3.3% 1.401e+10 perf-stat.i.dTLB-loads
1.015e+10 -3.5% 9.79e+09 perf-stat.i.dTLB-stores
4.808e+10 -3.3% 4.651e+10 perf-stat.i.instructions
0.79 -3.3% 0.77 perf-stat.i.ipc
20904568 -2.2% 20436868 perf-stat.i.node-loads
25841390 +79.8% 46459814 perf-stat.i.node-stores
45.31 -2.8% 44.03 perf-stat.overall.MPKI
13.35 +0.3 13.61 perf-stat.overall.cache-miss-rate%
1.26 +3.5% 1.31 perf-stat.overall.cpi
208.67 +4.4% 217.94 perf-stat.overall.cycles-between-cache-misses
0.79 -3.4% 0.77 perf-stat.overall.ipc
9.647e+09 -2.7% 9.387e+09 perf-stat.ps.branch-instructions
1.086e+08 -2.7% 1.056e+08 perf-stat.ps.branch-misses
2.903e+08 -4.1% 2.783e+08 perf-stat.ps.cache-misses
2.175e+09 -6.0% 2.045e+09 perf-stat.ps.cache-references
61620 -2.9% 59827 perf-stat.ps.cpu-migrations
1.447e+10 -3.3% 1.399e+10 perf-stat.ps.dTLB-loads
1.013e+10 -3.5% 9.774e+09 perf-stat.ps.dTLB-stores
4.8e+10 -3.3% 4.643e+10 perf-stat.ps.instructions
20870674 -2.2% 20403728 perf-stat.ps.node-loads
25799504 +79.8% 46384429 perf-stat.ps.node-stores
2.964e+13 -3.4% 2.863e+13 perf-stat.total.instructions
38.02 -1.3 36.70 perf-profile.calltrace.cycles-pp.ksys_read.__x64_sys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe
38.11 -1.3 36.80 perf-profile.calltrace.cycles-pp.__x64_sys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe
36.86 -1.3 35.58 perf-profile.calltrace.cycles-pp.vfs_read.ksys_read.__x64_sys_read.do_syscall_64.entry_SYSCALL_64_after_hwframe
34.13 -1.2 32.97 perf-profile.calltrace.cycles-pp.__vfs_read.vfs_read.ksys_read.__x64_sys_read.do_syscall_64
33.76 -1.1 32.61 perf-profile.calltrace.cycles-pp.new_sync_read.__vfs_read.vfs_read.ksys_read.__x64_sys_read
32.94 -1.1 31.84 perf-profile.calltrace.cycles-pp.sock_read_iter.new_sync_read.__vfs_read.vfs_read.ksys_read
31.73 -1.0 30.70 perf-profile.calltrace.cycles-pp.sock_recvmsg.sock_read_iter.new_sync_read.__vfs_read.vfs_read
31.32 -1.0 30.30 perf-profile.calltrace.cycles-pp.unix_stream_recvmsg.sock_recvmsg.sock_read_iter.new_sync_read.__vfs_read
30.92 -1.0 29.92 perf-profile.calltrace.cycles-pp.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg.sock_read_iter.new_sync_read
2.61 ± 3% -0.6 1.97 perf-profile.calltrace.cycles-pp.skb_set_owner_w.sock_alloc_send_pskb.unix_stream_sendmsg.sock_sendmsg.sock_write_iter
2.02 ± 4% -0.6 1.39 perf-profile.calltrace.cycles-pp.refcount_add_not_zero_checked.refcount_add_checked.skb_set_owner_w.sock_alloc_send_pskb.unix_stream_sendmsg
2.03 ± 4% -0.6 1.39 perf-profile.calltrace.cycles-pp.refcount_add_checked.skb_set_owner_w.sock_alloc_send_pskb.unix_stream_sendmsg.sock_sendmsg
2.68 -0.6 2.05 ± 2% perf-profile.calltrace.cycles-pp._raw_spin_lock.unix_stream_sendmsg.sock_sendmsg.sock_write_iter.new_sync_write
6.64 ± 3% -0.4 6.25 ± 2% perf-profile.calltrace.cycles-pp.sock_def_readable.unix_stream_sendmsg.sock_sendmsg.sock_write_iter.new_sync_write
9.06 -0.4 8.68 ± 2% perf-profile.calltrace.cycles-pp.unix_stream_read_actor.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg.sock_read_iter
8.98 -0.4 8.59 ± 2% perf-profile.calltrace.cycles-pp.skb_copy_datagram_iter.unix_stream_read_actor.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg
8.74 -0.4 8.36 ± 2% perf-profile.calltrace.cycles-pp.__skb_datagram_iter.skb_copy_datagram_iter.unix_stream_read_actor.unix_stream_read_generic.unix_stream_recvmsg
8.41 -0.4 8.05 ± 2% perf-profile.calltrace.cycles-pp.simple_copy_to_iter.__skb_datagram_iter.skb_copy_datagram_iter.unix_stream_read_actor.unix_stream_read_generic
11.82 -0.3 11.47 perf-profile.calltrace.cycles-pp.consume_skb.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg.sock_read_iter
6.11 -0.3 5.79 perf-profile.calltrace.cycles-pp._copy_to_iter.simple_copy_to_iter.__skb_datagram_iter.skb_copy_datagram_iter.unix_stream_read_actor
5.04 -0.3 4.79 ± 2% perf-profile.calltrace.cycles-pp.copy_user_enhanced_fast_string._copy_to_iter.simple_copy_to_iter.__skb_datagram_iter.skb_copy_datagram_iter
3.55 ± 3% -0.2 3.36 ± 2% perf-profile.calltrace.cycles-pp.__wake_up_common.__wake_up_common_lock.__wake_up_sync_key.sock_def_readable.unix_stream_sendmsg
3.57 -0.2 3.38 perf-profile.calltrace.cycles-pp.kmem_cache_free.kfree_skbmem.consume_skb.unix_stream_read_generic.unix_stream_recvmsg
7.34 -0.2 7.16 perf-profile.calltrace.cycles-pp.skb_release_all.consume_skb.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg
3.67 -0.2 3.50 perf-profile.calltrace.cycles-pp.kfree_skbmem.consume_skb.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg
3.47 -0.2 3.31 perf-profile.calltrace.cycles-pp.skb_release_head_state.skb_release_all.consume_skb.unix_stream_read_generic.unix_stream_recvmsg
2.90 -0.2 2.75 perf-profile.calltrace.cycles-pp.sock_wfree.unix_destruct_scm.skb_release_head_state.skb_release_all.consume_skb
1.85 ± 3% -0.1 1.70 ± 4% perf-profile.calltrace.cycles-pp._raw_spin_lock_irqsave.__wake_up_common_lock.__wake_up_sync_key.sock_def_readable.unix_stream_sendmsg
1.53 ± 3% -0.1 1.39 ± 4% perf-profile.calltrace.cycles-pp.syscall_return_via_sysret
2.32 ± 2% -0.1 2.18 perf-profile.calltrace.cycles-pp.___cache_free.kfree.skb_free_head.skb_release_data.skb_release_all
4.39 -0.1 4.25 perf-profile.calltrace.cycles-pp.skb_copy_datagram_from_iter.unix_stream_sendmsg.sock_sendmsg.sock_write_iter.new_sync_write
1.77 ± 3% -0.1 1.63 ± 4% perf-profile.calltrace.cycles-pp.queued_spin_lock_slowpath._raw_spin_lock_irqsave.__wake_up_common_lock.__wake_up_sync_key.sock_def_readable
3.28 -0.1 3.15 perf-profile.calltrace.cycles-pp.unix_destruct_scm.skb_release_head_state.skb_release_all.consume_skb.unix_stream_read_generic
2.04 ± 2% -0.1 1.92 ± 2% perf-profile.calltrace.cycles-pp.schedule_timeout.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg.sock_read_iter
2.67 -0.1 2.55 perf-profile.calltrace.cycles-pp.skb_queue_tail.unix_stream_sendmsg.sock_sendmsg.sock_write_iter.new_sync_write
1.96 ± 2% -0.1 1.84 ± 2% perf-profile.calltrace.cycles-pp.__sched_text_start.schedule.schedule_timeout.unix_stream_read_generic.unix_stream_recvmsg
2.00 ± 2% -0.1 1.88 ± 2% perf-profile.calltrace.cycles-pp.schedule.schedule_timeout.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg
2.68 ± 2% -0.1 2.57 ± 2% perf-profile.calltrace.cycles-pp._copy_from_iter.skb_copy_datagram_from_iter.unix_stream_sendmsg.sock_sendmsg.sock_write_iter
1.43 ± 2% -0.1 1.33 perf-profile.calltrace.cycles-pp.cache_alloc_refill.kmem_cache_alloc_node_trace.__kmalloc_node_track_caller.__kmalloc_reserve.__alloc_skb
1.01 ± 3% -0.1 0.91 ± 2% perf-profile.calltrace.cycles-pp._raw_spin_lock.___cache_free.kfree.skb_free_head.skb_release_data
3.91 -0.1 3.81 perf-profile.calltrace.cycles-pp.syscall_slow_exit_work.do_syscall_64.entry_SYSCALL_64_after_hwframe
1.40 ± 4% -0.1 1.29 ± 2% perf-profile.calltrace.cycles-pp.skb_unlink.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg.sock_read_iter
1.71 ± 2% -0.1 1.61 perf-profile.calltrace.cycles-pp.___cache_free.kmem_cache_free.kfree_skbmem.consume_skb.unix_stream_read_generic
0.93 ± 3% -0.1 0.83 ± 2% perf-profile.calltrace.cycles-pp.queued_spin_lock_slowpath._raw_spin_lock.___cache_free.kfree.skb_free_head
3.24 -0.1 3.15 perf-profile.calltrace.cycles-pp.__audit_syscall_exit.syscall_slow_exit_work.do_syscall_64.entry_SYSCALL_64_after_hwframe
1.38 ± 2% -0.1 1.29 perf-profile.calltrace.cycles-pp.rw_verify_area.vfs_read.ksys_read.__x64_sys_read.do_syscall_64
3.74 -0.1 3.66 perf-profile.calltrace.cycles-pp.kmem_cache_alloc_node.__alloc_skb.alloc_skb_with_frags.sock_alloc_send_pskb.unix_stream_sendmsg
0.96 ± 4% -0.1 0.89 ± 2% perf-profile.calltrace.cycles-pp.stack_trace_save_tsk.__account_scheduler_latency.enqueue_entity.enqueue_task_fair.activate_task
0.67 ± 3% -0.1 0.60 perf-profile.calltrace.cycles-pp._raw_spin_lock.cache_alloc_refill.kmem_cache_alloc_node_trace.__kmalloc_node_track_caller.__kmalloc_reserve
0.97 -0.1 0.90 perf-profile.calltrace.cycles-pp.security_file_permission.rw_verify_area.vfs_read.ksys_read.__x64_sys_read
0.61 ± 3% -0.1 0.55 perf-profile.calltrace.cycles-pp.queued_spin_lock_slowpath._raw_spin_lock.cache_alloc_refill.kmem_cache_alloc_node_trace.__kmalloc_node_track_caller
0.81 ± 4% -0.1 0.75 perf-profile.calltrace.cycles-pp.arch_stack_walk.stack_trace_save_tsk.__account_scheduler_latency.enqueue_entity.enqueue_task_fair
0.57 ± 5% -0.0 0.52 ± 2% perf-profile.calltrace.cycles-pp._raw_spin_unlock_irqrestore.skb_queue_tail.unix_stream_sendmsg.sock_sendmsg.sock_write_iter
0.61 ± 2% -0.0 0.57 ± 3% perf-profile.calltrace.cycles-pp._raw_spin_lock.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg.sock_read_iter
0.66 -0.0 0.61 ± 3% perf-profile.calltrace.cycles-pp._raw_spin_lock_irqsave.skb_unlink.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg
0.65 -0.0 0.61 ± 4% perf-profile.calltrace.cycles-pp.wait_for_unix_gc.unix_stream_sendmsg.sock_sendmsg.sock_write_iter.new_sync_write
1.13 ± 2% +0.1 1.23 perf-profile.calltrace.cycles-pp.ksize.__alloc_skb.alloc_skb_with_frags.sock_alloc_send_pskb.unix_stream_sendmsg
1.10 +0.1 1.21 perf-profile.calltrace.cycles-pp.__ksize.ksize.__alloc_skb.alloc_skb_with_frags.sock_alloc_send_pskb
0.38 ± 57% +0.1 0.53 perf-profile.calltrace.cycles-pp.refcount_dec_and_test_checked.consume_skb.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg
1.80 ± 2% +0.2 1.96 perf-profile.calltrace.cycles-pp.refcount_inc_not_zero_checked.unix_stream_read_generic.unix_stream_recvmsg.sock_recvmsg.sock_read_iter
92.56 +0.4 92.96 perf-profile.calltrace.cycles-pp.entry_SYSCALL_64_after_hwframe
90.14 +0.7 90.81 perf-profile.calltrace.cycles-pp.do_syscall_64.entry_SYSCALL_64_after_hwframe
0.00 +0.9 0.86 ± 6% perf-profile.calltrace.cycles-pp.memset_erms.__alloc_skb.alloc_skb_with_frags.sock_alloc_send_pskb.unix_stream_sendmsg
0.00 +0.9 0.89 ± 8% perf-profile.calltrace.cycles-pp.memset_erms.__kmalloc_node_track_caller.__kmalloc_reserve.__alloc_skb.alloc_skb_with_frags
4.35 +1.5 5.85 perf-profile.calltrace.cycles-pp.kmem_cache_alloc_node_trace.__kmalloc_node_track_caller.__kmalloc_reserve.__alloc_skb.alloc_skb_with_frags
44.05 +2.2 46.24 perf-profile.calltrace.cycles-pp.ksys_write.__x64_sys_write.do_syscall_64.entry_SYSCALL_64_after_hwframe
44.14 +2.2 46.34 perf-profile.calltrace.cycles-pp.__x64_sys_write.do_syscall_64.entry_SYSCALL_64_after_hwframe
41.66 +2.3 43.94 perf-profile.calltrace.cycles-pp.vfs_write.ksys_write.__x64_sys_write.do_syscall_64.entry_SYSCALL_64_after_hwframe
39.46 +2.4 41.82 perf-profile.calltrace.cycles-pp.__vfs_write.vfs_write.ksys_write.__x64_sys_write.do_syscall_64
39.21 +2.4 41.58 perf-profile.calltrace.cycles-pp.new_sync_write.__vfs_write.vfs_write.ksys_write.__x64_sys_write
38.46 +2.4 40.85 perf-profile.calltrace.cycles-pp.sock_write_iter.new_sync_write.__vfs_write.vfs_write.ksys_write
4.91 +2.4 7.30 perf-profile.calltrace.cycles-pp.__kmalloc_node_track_caller.__kmalloc_reserve.__alloc_skb.alloc_skb_with_frags.sock_alloc_send_pskb
37.20 +2.4 39.63 perf-profile.calltrace.cycles-pp.sock_sendmsg.sock_write_iter.new_sync_write.__vfs_write.vfs_write
36.36 +2.4 38.80 perf-profile.calltrace.cycles-pp.unix_stream_sendmsg.sock_sendmsg.sock_write_iter.new_sync_write.__vfs_write
5.15 +2.6 7.74 perf-profile.calltrace.cycles-pp.__kmalloc_reserve.__alloc_skb.alloc_skb_with_frags.sock_alloc_send_pskb.unix_stream_sendmsg
16.15 +3.9 20.05 perf-profile.calltrace.cycles-pp.sock_alloc_send_pskb.unix_stream_sendmsg.sock_sendmsg.sock_write_iter.new_sync_write
11.73 +4.5 16.21 perf-profile.calltrace.cycles-pp.__alloc_skb.alloc_skb_with_frags.sock_alloc_send_pskb.unix_stream_sendmsg.sock_sendmsg
12.06 +4.5 16.55 perf-profile.calltrace.cycles-pp.alloc_skb_with_frags.sock_alloc_send_pskb.unix_stream_sendmsg.sock_sendmsg.sock_write_iter
38.39 -1.3 37.06 perf-profile.children.cycles-pp.__x64_sys_read
38.13 -1.3 36.81 perf-profile.children.cycles-pp.ksys_read
36.94 -1.3 35.66 perf-profile.children.cycles-pp.vfs_read
34.24 -1.2 33.07 perf-profile.children.cycles-pp.__vfs_read
33.84 -1.2 32.68 perf-profile.children.cycles-pp.new_sync_read
33.17 -1.1 32.06 perf-profile.children.cycles-pp.sock_read_iter
31.60 -1.0 30.55 perf-profile.children.cycles-pp.unix_stream_recvmsg
31.85 -1.0 30.81 perf-profile.children.cycles-pp.sock_recvmsg
31.02 -1.0 30.02 perf-profile.children.cycles-pp.unix_stream_read_generic
6.85 -1.0 5.88 perf-profile.children.cycles-pp._raw_spin_lock
2.65 ± 3% -0.6 2.00 perf-profile.children.cycles-pp.skb_set_owner_w
2.08 ± 4% -0.6 1.44 perf-profile.children.cycles-pp.refcount_add_checked
2.23 ± 4% -0.6 1.62 perf-profile.children.cycles-pp.refcount_add_not_zero_checked
6.85 ± 2% -0.4 6.44 ± 2% perf-profile.children.cycles-pp.sock_def_readable
9.10 -0.4 8.70 ± 2% perf-profile.children.cycles-pp.unix_stream_read_actor
9.04 -0.4 8.65 ± 2% perf-profile.children.cycles-pp.skb_copy_datagram_iter
8.79 -0.4 8.41 ± 2% perf-profile.children.cycles-pp.__skb_datagram_iter
8.50 -0.4 8.14 ± 2% perf-profile.children.cycles-pp.simple_copy_to_iter
11.93 -0.3 11.59 perf-profile.children.cycles-pp.consume_skb
5.62 -0.3 5.29 perf-profile.children.cycles-pp.queued_spin_lock_slowpath
6.17 -0.3 5.84 perf-profile.children.cycles-pp._copy_to_iter
6.71 -0.3 6.42 perf-profile.children.cycles-pp.copy_user_enhanced_fast_string
4.08 -0.2 3.84 perf-profile.children.cycles-pp.___cache_free
3.02 -0.2 2.80 ± 2% perf-profile.children.cycles-pp.syscall_return_via_sysret
5.15 -0.2 4.93 perf-profile.children.cycles-pp._raw_spin_lock_irqsave
3.66 -0.2 3.46 perf-profile.children.cycles-pp.kmem_cache_free
7.41 -0.2 7.22 perf-profile.children.cycles-pp.skb_release_all
3.74 -0.2 3.58 perf-profile.children.cycles-pp.kfree_skbmem
3.01 -0.2 2.85 perf-profile.children.cycles-pp.sock_wfree
3.51 -0.2 3.36 perf-profile.children.cycles-pp.skb_release_head_state
2.38 ± 2% -0.1 2.23 perf-profile.children.cycles-pp.__might_sleep
4.44 -0.1 4.30 perf-profile.children.cycles-pp.skb_copy_datagram_from_iter
3.53 -0.1 3.40 perf-profile.children.cycles-pp.unix_destruct_scm
1.55 ± 3% -0.1 1.42 ± 2% perf-profile.children.cycles-pp.___might_sleep
2.10 ± 2% -0.1 1.97 perf-profile.children.cycles-pp.cache_alloc_refill
2.70 -0.1 2.58 perf-profile.children.cycles-pp.skb_queue_tail
2.75 ± 2% -0.1 2.64 ± 2% perf-profile.children.cycles-pp._copy_from_iter
1.84 -0.1 1.74 perf-profile.children.cycles-pp.rw_verify_area
4.03 -0.1 3.93 perf-profile.children.cycles-pp.syscall_slow_exit_work
1.06 ± 4% -0.1 0.96 ± 3% perf-profile.children.cycles-pp.unwind_next_frame
1.41 ± 4% -0.1 1.31 ± 2% perf-profile.children.cycles-pp.skb_unlink
2.71 -0.1 2.61 perf-profile.children.cycles-pp.schedule_timeout
2.83 -0.1 2.73 perf-profile.children.cycles-pp.__sched_text_start
3.43 -0.1 3.34 perf-profile.children.cycles-pp.__audit_syscall_exit
2.65 -0.1 2.57 perf-profile.children.cycles-pp.schedule
1.22 ± 4% -0.1 1.14 perf-profile.children.cycles-pp.stack_trace_save_tsk
1.15 -0.1 1.07 perf-profile.children.cycles-pp.security_file_permission
0.58 ± 5% -0.1 0.50 ± 3% perf-profile.children.cycles-pp.memcg_kmem_put_cache
2.65 -0.1 2.58 ± 2% perf-profile.children.cycles-pp.__fdget_pos
1.06 ± 3% -0.1 0.99 perf-profile.children.cycles-pp.arch_stack_walk
0.90 ± 3% -0.1 0.83 ± 3% perf-profile.children.cycles-pp.__might_fault
1.42 ± 2% -0.1 1.37 perf-profile.children.cycles-pp.__fsnotify_parent
0.17 ± 7% -0.1 0.12 ± 10% perf-profile.children.cycles-pp.should_failslab
1.03 -0.1 0.98 perf-profile.children.cycles-pp.__list_del_entry_valid
0.89 ± 2% -0.0 0.85 ± 2% perf-profile.children.cycles-pp.deactivate_task
0.30 -0.0 0.26 ± 4% perf-profile.children.cycles-pp.copyin
1.07 ± 2% -0.0 1.04 perf-profile.children.cycles-pp.unroll_tree_refs
1.01 -0.0 0.98 perf-profile.children.cycles-pp.mutex_lock
0.35 ± 2% -0.0 0.32 ± 5% perf-profile.children.cycles-pp.dequeue_entity
0.36 ± 3% -0.0 0.34 ± 2% perf-profile.children.cycles-pp.select_idle_sibling
0.25 ± 2% -0.0 0.23 ± 2% perf-profile.children.cycles-pp.copyout
0.33 ± 3% -0.0 0.31 ± 3% perf-profile.children.cycles-pp.path_put
0.56 +0.0 0.59 ± 2% perf-profile.children.cycles-pp.refcount_dec_and_test_checked
0.08 ± 10% +0.0 0.11 ± 7% perf-profile.children.cycles-pp.cache_grow_begin
0.11 ± 6% +0.0 0.15 ± 14% perf-profile.children.cycles-pp.kmalloc_slab
0.00 +0.1 0.08 ± 10% perf-profile.children.cycles-pp.memset
1.15 +0.1 1.26 perf-profile.children.cycles-pp.__ksize
1.18 ± 2% +0.1 1.30 perf-profile.children.cycles-pp.ksize
2.52 ± 2% +0.2 2.69 perf-profile.children.cycles-pp.refcount_inc_not_zero_checked
92.60 +0.4 93.00 perf-profile.children.cycles-pp.entry_SYSCALL_64_after_hwframe
90.32 +0.7 90.98 perf-profile.children.cycles-pp.do_syscall_64
4.57 +1.6 6.21 perf-profile.children.cycles-pp.kmem_cache_alloc_node_trace
0.00 +1.8 1.77 ± 8% perf-profile.children.cycles-pp.memset_erms
44.45 +2.2 46.62 perf-profile.children.cycles-pp.__x64_sys_write
44.16 +2.2 46.36 perf-profile.children.cycles-pp.ksys_write
41.70 +2.3 43.98 perf-profile.children.cycles-pp.vfs_write
39.54 +2.4 41.89 perf-profile.children.cycles-pp.__vfs_write
39.29 +2.4 41.66 perf-profile.children.cycles-pp.new_sync_write
38.63 +2.4 41.02 perf-profile.children.cycles-pp.sock_write_iter
4.98 +2.4 7.41 perf-profile.children.cycles-pp.__kmalloc_node_track_caller
37.26 +2.4 39.70 perf-profile.children.cycles-pp.sock_sendmsg
36.68 +2.4 39.13 perf-profile.children.cycles-pp.unix_stream_sendmsg
5.18 +2.6 7.80 perf-profile.children.cycles-pp.__kmalloc_reserve
16.19 +3.9 20.09 perf-profile.children.cycles-pp.sock_alloc_send_pskb
11.81 +4.5 16.28 perf-profile.children.cycles-pp.__alloc_skb
12.11 +4.5 16.59 perf-profile.children.cycles-pp.alloc_skb_with_frags
3.37 -0.7 2.63 perf-profile.self.cycles-pp._raw_spin_lock
2.21 ± 3% -0.6 1.60 perf-profile.self.cycles-pp.refcount_add_not_zero_checked
5.60 -0.3 5.27 perf-profile.self.cycles-pp.queued_spin_lock_slowpath
6.68 -0.3 6.39 perf-profile.self.cycles-pp.copy_user_enhanced_fast_string
1.77 -0.2 1.55 perf-profile.self.cycles-pp.entry_SYSCALL_64_after_hwframe
2.98 -0.2 2.77 ± 2% perf-profile.self.cycles-pp.syscall_return_via_sysret
1.83 ± 4% -0.2 1.65 ± 6% perf-profile.self.cycles-pp.__check_object_size
1.45 ± 3% -0.1 1.34 ± 2% perf-profile.self.cycles-pp.___might_sleep
2.98 -0.1 2.88 perf-profile.self.cycles-pp._raw_spin_lock_irqsave
0.65 -0.1 0.55 ± 2% perf-profile.self.cycles-pp.sock_wfree
1.71 -0.1 1.61 ± 2% perf-profile.self.cycles-pp.unix_stream_read_generic
2.94 -0.1 2.85 perf-profile.self.cycles-pp.unix_stream_sendmsg
1.85 -0.1 1.78 perf-profile.self.cycles-pp.kmem_cache_free
0.92 ± 3% -0.1 0.85 ± 4% perf-profile.self.cycles-pp.__audit_syscall_entry
1.00 ± 2% -0.1 0.94 perf-profile.self.cycles-pp.__list_del_entry_valid
1.10 ± 3% -0.1 1.05 ± 2% perf-profile.self.cycles-pp.sock_read_iter
0.50 ± 4% -0.1 0.45 ± 4% perf-profile.self.cycles-pp.memcg_kmem_put_cache
0.71 ± 3% -0.1 0.66 perf-profile.self.cycles-pp.new_sync_read
1.00 ± 2% -0.0 0.95 ± 3% perf-profile.self.cycles-pp.__might_sleep
1.83 -0.0 1.79 perf-profile.self.cycles-pp.__audit_syscall_exit
0.34 ± 5% -0.0 0.30 ± 4% perf-profile.self.cycles-pp.skb_copy_datagram_from_iter
0.18 ± 4% -0.0 0.14 ± 5% perf-profile.self.cycles-pp.copyin
0.40 ± 2% -0.0 0.37 ± 5% perf-profile.self.cycles-pp.skb_set_owner_w
0.10 ± 8% -0.0 0.08 ± 5% perf-profile.self.cycles-pp.should_failslab
0.23 ± 3% -0.0 0.21 ± 3% perf-profile.self.cycles-pp.skb_copy_datagram_iter
0.13 ± 6% -0.0 0.11 ± 7% perf-profile.self.cycles-pp.update_rq_clock
0.14 ± 6% -0.0 0.12 ± 3% perf-profile.self.cycles-pp.stack_trace_consume_entry_nosched
0.11 ± 4% -0.0 0.10 ± 4% perf-profile.self.cycles-pp.__update_load_avg_se
0.11 +0.0 0.12 ± 4% perf-profile.self.cycles-pp.kfree_skbmem
0.31 ± 2% +0.0 0.34 ± 2% perf-profile.self.cycles-pp.skb_release_data
0.10 ± 11% +0.0 0.13 ± 14% perf-profile.self.cycles-pp.kmalloc_slab
0.18 ± 2% +0.0 0.23 ± 4% perf-profile.self.cycles-pp.__kmalloc_reserve
0.15 ± 3% +0.1 0.21 ± 2% perf-profile.self.cycles-pp.__kmalloc_node_track_caller
1.07 ± 2% +0.1 1.18 ± 2% perf-profile.self.cycles-pp.kfree
1.12 +0.1 1.24 perf-profile.self.cycles-pp.__ksize
2.51 ± 2% +0.2 2.67 perf-profile.self.cycles-pp.refcount_inc_not_zero_checked
1.38 ± 10% +1.0 2.39 perf-profile.self.cycles-pp.__alloc_skb
0.00 +1.7 1.73 ± 8% perf-profile.self.cycles-pp.memset_erms
1.26 +1.8 3.08 perf-profile.self.cycles-pp.kmem_cache_alloc_node_trace
Disclaimer:
Results have been estimated based on internal Intel analysis and are provided
for informational purposes only. Any difference in system hardware or software
design or configuration may affect actual performance.
Thanks,
Oliver Sang
View attachment "config-5.2.0-05754-g6471384af2a65" of type "text/plain" (200966 bytes)
View attachment "job-script" of type "text/plain" (7232 bytes)
View attachment "job.yaml" of type "text/plain" (4851 bytes)
View attachment "reproduce" of type "text/plain" (3026 bytes)
Powered by blists - more mailing lists