Content-Type: multipart/mixed; boundary="===============0663821624838511493==" MIME-Version: 1.0 From: Qu Wenruo To: lkp@lists.01.org Subject: Re: [btrfs] e86bb85b1f: stress-ng.utime.ops_per_sec -70.1% regression Date: Wed, 13 Jan 2021 15:15:09 +0800 Message-ID: In-Reply-To: <20210112152415.GG30747@xsang-OptiPlex-9020> List-Id: --===============0663821624838511493== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable On 2021/1/12 =E4=B8=8B=E5=8D=8811:24, kernel test robot wrote: > > Greeting, > > FYI, we noticed a -70.1% regression of stress-ng.utime.ops_per_sec due to= commit: > > > commit: e86bb85b1fec48bcb8dfb79ec9f104d1a38fda78 ("[PATCH] btrfs: make bt= rfs_dirty_inode() to always reserve metadata space") > url: https://github.com/0day-ci/linux/commits/Qu-Wenruo/btrfs-make-btrfs_= dirty_inode-to-always-reserve-metadata-space/20210108-134133 > base: https://git.kernel.org/cgit/linux/kernel/git/kdave/linux.git for-ne= xt > > in testcase: stress-ng > on test machine: 96 threads Intel(R) Xeon(R) Gold 6252 CPU @ 2.10GHz with= 512G memory > with following parameters: > > nr_threads: 10% > disk: 1HDD > testtime: 30s > class: filesystem > cpufreq_governor: performance > ucode: 0x5003003 > fs: btrfs > > > > > If you fix the issue, kindly add following tag > Reported-by: kernel test robot > > > 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 > > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D > class/compiler/cpufreq_governor/disk/fs/kconfig/nr_threads/rootfs/tbox_gr= oup/testcase/testtime/ucode: > filesystem/gcc-9/performance/1HDD/btrfs/x86_64-rhel-8.3/10%/debian-10.= 4-x86_64-20200603.cgz/lkp-csl-2sp7/stress-ng/30s/0x5003003 > > commit: > 97847e0652 ("Merge branch 'for-next-next-v5.10-20201211' into for-next= -20201211") > e86bb85b1f ("btrfs: make btrfs_dirty_inode() to always reserve metadat= a space") > > 97847e06525b51ea e86bb85b1fec48bcb8dfb79ec9f > ---------------- --------------------------- > %stddev %change %stddev > \ | \ > 1098218 -40.4% 654054 stress-ng.access.ops > 36607 -40.4% 21801 stress-ng.access.ops_per_s= ec This is a little interesting. Although accessing an inode will update its atime, but don't we have lazy_atime mount option? > 92962 =C2=B1 2% -44.1% 51992 =C2=B1 3% stress-ng.chmod.= ops > 3098 =C2=B1 2% -44.1% 1733 =C2=B1 3% stress-ng.chmod.= ops_per_sec > 936128 =C2=B1 6% -41.0% 552284 stress-ng.chown.ops > 31204 =C2=B1 6% -41.0% 18409 stress-ng.chown.ops_p= er_sec > 1939514 -18.5% 1580533 stress-ng.fcntl.ops > 64650 -18.5% 52684 stress-ng.fcntl.ops_per_sec > 3705607 =C2=B1 2% -70.1% 1109769 stress-ng.utime.ops > 123519 =C2=B1 2% -70.1% 36992 stress-ng.utime.ops_p= er_sec Another interesting part is, only stress-ng is reporting such regressioin on the commit? No other report on the commit with different test env? E.g. NVME SSD? Above operations is affected by such commit, but I'm a little surprised only one report here. Just because flushing on HDD is more expensive? If no other test suite is fine, I would prefer to accept the drop, as it really streamline the operations. Thanks, Qu > 381.20 =C2=B1 6% +12.3% 428.27 =C2=B1 9% sched_debug.cfs_= rq:/.load_avg.avg > 6316 =C2=B1 57% -79.8% 1278 =C2=B1 68% softirqs.CPU77.B= LOCK > 10488 =C2=B1101% -89.5% 1100 =C2=B1124% softirqs.CPU78.B= LOCK > 5605 =C2=B1 92% -82.3% 990.50 =C2=B1 32% softirqs.CPU80.B= LOCK > 6094 =C2=B1128% -89.9% 614.50 =C2=B1 44% softirqs.CPU92.B= LOCK > 4921 =C2=B1 9% +20.5% 5931 =C2=B1 5% slabinfo.dmaengi= ne-unmap-16.active_objs > 4922 =C2=B1 9% +20.5% 5933 =C2=B1 5% slabinfo.dmaengi= ne-unmap-16.num_objs > 9818 =C2=B1 5% -6.9% 9139 =C2=B1 3% slabinfo.kmalloc= -rcl-256.active_objs > 49223 =C2=B1 3% -18.4% 40177 =C2=B1 3% slabinfo.radix_t= ree_node.active_objs > 903.25 =C2=B1 3% -18.0% 740.50 =C2=B1 3% slabinfo.radix_t= ree_node.active_slabs > 50620 =C2=B1 3% -18.0% 41505 =C2=B1 3% slabinfo.radix_t= ree_node.num_objs > 903.25 =C2=B1 3% -18.0% 740.50 =C2=B1 3% slabinfo.radix_t= ree_node.num_slabs > 9927 =C2=B1 3% +5.8% 10504 proc-vmstat.nr_active= _anon > 6043459 =C2=B1 2% -2.2% 5911900 proc-vmstat.nr_dirtied > 1125 -6.1% 1056 =C2=B1 4% proc-vmstat.nr_dirty > 20361 =C2=B1 2% +4.7% 21309 proc-vmstat.nr_shmem > 66221 -4.3% 63404 =C2=B1 2% proc-vmstat.nr_slab_r= eclaimable > 9927 =C2=B1 3% +5.8% 10504 proc-vmstat.nr_zone_a= ctive_anon > 1225 -5.8% 1154 =C2=B1 3% proc-vmstat.nr_zone_w= rite_pending > 11313111 -2.1% 11072335 proc-vmstat.pgfault > 0.00 +125.0% 0.00 =C2=B1 19% perf-sched.sch_delay.= avg.ms.preempt_schedule_common._cond_resched.kmem_cache_alloc.start_transac= tion.btrfs_dirty_inode > 0.01 =C2=B1 13% -24.5% 0.01 =C2=B1 15% perf-sched.sch_d= elay.max.ms.do_wait.kernel_wait4.__do_sys_wait4.do_syscall_64 > 3.26 =C2=B1 41% -98.8% 0.04 perf-sched.sch_delay.= max.ms.rwsem_down_write_slowpath.chmod_common.do_fchmodat.__x64_sys_chmod > 0.00 =C2=B1 50% +123.5% 0.01 =C2=B1 24% perf-sched.sch_d= elay.max.ms.wait_for_partner.fifo_open.do_dentry_open.path_openat > 4.77 =C2=B1 6% -25.7% 3.54 perf-sched.total_sch_= delay.max.ms > 0.02 =C2=B1 7% +19.5% 0.02 =C2=B1 5% perf-sched.wait_= and_delay.avg.ms.rwsem_down_write_slowpath.chmod_common.do_fchmodat.__x64_s= ys_fchmodat > 1711 =C2=B1 27% +60.7% 2750 =C2=B1 22% perf-sched.wait_= and_delay.count.rwsem_down_write_slowpath.chmod_common.__x64_sys_fchmod.do_= syscall_64 > 0.03 =C2=B1132% +2497.6% 0.81 =C2=B1 66% perf-sched.wait_= time.avg.ms.preempt_schedule_common._cond_resched.dput.path_put.do_fchmodat > 0.25 =C2=B1 67% +243.4% 0.86 =C2=B1 49% perf-sched.wait_= time.avg.ms.preempt_schedule_common._cond_resched.kmem_cache_alloc.start_tr= ansaction.btrfs_dirty_inode > 0.02 =C2=B1 11% +28.6% 0.02 =C2=B1 13% perf-sched.wait_= time.avg.ms.rwsem_down_write_slowpath.chmod_common.do_fchmodat.__x64_sys_ch= mod > 0.27 =C2=B1160% +1248.5% 3.67 =C2=B1 57% perf-sched.wait_= time.max.ms.preempt_schedule_common._cond_resched.dput.path_put.do_fchmodat > 0.04 =C2=B1165% +4851.4% 2.17 =C2=B1 99% perf-sched.wait_= time.max.ms.preempt_schedule_common._cond_resched.mnt_want_write.chmod_comm= on.do_fchmodat > 36969238 -5.8% 34829162 =C2=B1 2% perf-stat.i.cache-mis= ses > 2.009e+09 -2.7% 1.954e+09 perf-stat.i.dTLB-stores > 12297175 =C2=B1 2% -6.5% 11500340 perf-stat.i.iTLB-load= -misses > 1843 +7.0% 1972 =C2=B1 2% perf-stat.i.instructi= ons-per-iTLB-miss > 8936 -2.2% 8740 perf-stat.i.minor-faults > 952962 -3.6% 919068 perf-stat.i.node-loads > 5348157 -7.4% 4950006 =C2=B1 2% perf-stat.i.node-stor= e-misses > 8942 -2.2% 8748 perf-stat.i.page-faults > 0.00 =C2=B1 20% +0.0 0.00 =C2=B1 18% perf-stat.overal= l.dTLB-store-miss-rate% > 1263 +6.1% 1339 perf-stat.overall.instruct= ions-per-iTLB-miss > 36924576 -5.8% 34784593 =C2=B1 2% perf-stat.ps.cache-mi= sses > 2.003e+09 -2.7% 1.949e+09 perf-stat.ps.dTLB-stores > 12274047 =C2=B1 2% -6.5% 11479296 perf-stat.ps.iTLB-loa= d-misses > 8880 -2.1% 8694 perf-stat.ps.minor-faults > 950813 -3.6% 916929 perf-stat.ps.node-loads > 5343222 -7.5% 4944422 =C2=B1 2% perf-stat.ps.node-sto= re-misses > 8887 -2.1% 8702 perf-stat.ps.page-faults > 57.00 =C2=B1166% -100.0% 0.00 interrupts.92:PCI-MSI= .31981625-edge.i40e-eth0-TxRx-56 > 19339 =C2=B1 7% +31.3% 25385 =C2=B1 16% interrupts.CPU13= .CAL:Function_call_interrupts > 19917 =C2=B1 5% +9.7% 21856 interrupts.CPU23.TLB:= TLB_shootdowns > 20940 =C2=B1 5% +9.5% 22935 =C2=B1 4% interrupts.CPU29= .CAL:Function_call_interrupts > 3470 =C2=B1 65% -96.1% 136.75 =C2=B1 5% interrupts.CPU29= .NMI:Non-maskable_interrupts > 3470 =C2=B1 65% -96.1% 136.75 =C2=B1 5% interrupts.CPU29= .PMI:Performance_monitoring_interrupts > 2061 =C2=B1160% -93.5% 134.00 =C2=B1 3% interrupts.CPU33= .NMI:Non-maskable_interrupts > 2061 =C2=B1160% -93.5% 134.00 =C2=B1 3% interrupts.CPU33= .PMI:Performance_monitoring_interrupts > 3041 =C2=B1 47% +75.6% 5342 =C2=B1 36% interrupts.CPU35= .RES:Rescheduling_interrupts > 3748 =C2=B1 96% -96.5% 133.00 =C2=B1 5% interrupts.CPU40= .NMI:Non-maskable_interrupts > 3748 =C2=B1 96% -96.5% 133.00 =C2=B1 5% interrupts.CPU40= .PMI:Performance_monitoring_interrupts > 21172 =C2=B1 5% -12.7% 18476 =C2=B1 6% interrupts.CPU45= .CAL:Function_call_interrupts > 56.25 =C2=B1169% -100.0% 0.00 interrupts.CPU56.92:P= CI-MSI.31981625-edge.i40e-eth0-TxRx-56 > 122.50 =C2=B1 25% -34.5% 80.25 =C2=B1 32% interrupts.CPU56= .NMI:Non-maskable_interrupts > 122.50 =C2=B1 25% -34.5% 80.25 =C2=B1 32% interrupts.CPU56= .PMI:Performance_monitoring_interrupts > 21585 =C2=B1 5% +13.4% 24479 =C2=B1 5% interrupts.CPU58= .TLB:TLB_shootdowns > 3340 =C2=B1 22% +76.3% 5888 =C2=B1 48% interrupts.CPU62= .RES:Rescheduling_interrupts > 23969 =C2=B1 5% +12.5% 26976 =C2=B1 14% interrupts.CPU66= .CAL:Function_call_interrupts > 22832 =C2=B1 5% +26.5% 28873 =C2=B1 14% interrupts.CPU69= .CAL:Function_call_interrupts > 2329 =C2=B1106% -66.3% 784.50 =C2=B1150% interrupts.CPU73= .NMI:Non-maskable_interrupts > 2329 =C2=B1106% -66.3% 784.50 =C2=B1150% interrupts.CPU73= .PMI:Performance_monitoring_interrupts > 3520 =C2=B1 65% -96.7% 116.00 =C2=B1 27% interrupts.CPU77= .NMI:Non-maskable_interrupts > 3520 =C2=B1 65% -96.7% 116.00 =C2=B1 27% interrupts.CPU77= .PMI:Performance_monitoring_interrupts > 1570 =C2=B1156% -93.6% 101.00 =C2=B1 37% interrupts.CPU81= .NMI:Non-maskable_interrupts > 1570 =C2=B1156% -93.6% 101.00 =C2=B1 37% interrupts.CPU81= .PMI:Performance_monitoring_interrupts > 2501 =C2=B1109% -95.9% 102.75 =C2=B1 41% interrupts.CPU88= .NMI:Non-maskable_interrupts > 2501 =C2=B1109% -95.9% 102.75 =C2=B1 41% interrupts.CPU88= .PMI:Performance_monitoring_interrupts > 5208 =C2=B1 44% -48.6% 2674 =C2=B1 46% interrupts.CPU9.= RES:Rescheduling_interrupts > 20273 =C2=B1 10% +22.3% 24803 =C2=B1 11% interrupts.CPU91= .CAL:Function_call_interrupts > 1833 =C2=B1158% -93.6% 117.75 =C2=B1 24% interrupts.CPU94= .NMI:Non-maskable_interrupts > 1833 =C2=B1158% -93.6% 117.75 =C2=B1 24% interrupts.CPU94= .PMI:Performance_monitoring_interrupts > 47.69 =C2=B1 11% -27.3 20.39 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.do_faccessat.do_syscall_64.entry_SYSCALL_64_after_hwframe > 20.56 =C2=B1 11% -12.1 8.50 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.prepare_creds.do_faccessat.do_syscall_64.entry_SYSCALL_64_= after_hwframe > 18.78 =C2=B1 11% -11.9 6.92 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.put_cred_rcu.do_faccessat.do_syscall_64.entry_SYSCALL_64_a= fter_hwframe > 5.42 =C2=B1 12% -3.8 1.60 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.free_uid.put_cred_rcu.do_faccessat.do_syscall_64.entry_SYS= CALL_64_after_hwframe > 5.38 =C2=B1 12% -3.8 1.57 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.refcount_dec_and_lock_irqsave.free_uid.put_cred_rcu.do_fac= cessat.do_syscall_64 > 5.36 =C2=B1 12% -3.8 1.56 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.refcount_dec_not_one.refcount_dec_and_lock_irqsave.free_ui= d.put_cred_rcu.do_faccessat > 5.32 =C2=B1 11% -3.1 2.19 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.key_put.put_cred_rcu.do_faccessat.do_syscall_64.entry_SYSC= ALL_64_after_hwframe > 5.17 =C2=B1 11% -2.4 2.74 =C2=B1 4% perf-profile.cal= ltrace.cycles-pp.security_prepare_creds.prepare_creds.do_faccessat.do_sysca= ll_64.entry_SYSCALL_64_after_hwframe > 4.77 =C2=B1 8% -2.3 2.50 =C2=B1 4% perf-profile.cal= ltrace.cycles-pp.btrfs_update_inode.btrfs_dirty_inode.btrfs_setattr.notify_= change.chmod_common > 3.79 =C2=B1 7% -2.0 1.84 =C2=B1 4% perf-profile.cal= ltrace.cycles-pp.btrfs_delayed_update_inode.btrfs_update_inode.btrfs_dirty_= inode.btrfs_setattr.notify_change > 3.54 =C2=B1 12% -1.8 1.74 =C2=B1 4% perf-profile.cal= ltrace.cycles-pp.apparmor_cred_prepare.security_prepare_creds.prepare_creds= .do_faccessat.do_syscall_64 > 3.49 =C2=B1 9% -1.8 1.72 =C2=B1 6% perf-profile.cal= ltrace.cycles-pp.security_cred_free.put_cred_rcu.do_faccessat.do_syscall_64= .entry_SYSCALL_64_after_hwframe > 2.49 =C2=B1 9% -1.4 1.08 =C2=B1 8% perf-profile.cal= ltrace.cycles-pp.apparmor_cred_free.security_cred_free.put_cred_rcu.do_facc= essat.do_syscall_64 > 3.38 =C2=B1 11% -1.3 2.04 =C2=B1 6% perf-profile.cal= ltrace.cycles-pp.filename_lookup.do_faccessat.do_syscall_64.entry_SYSCALL_6= 4_after_hwframe > 3.25 =C2=B1 11% -1.3 1.96 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.__do_sys_newfstatat.do_syscall_64.entry_SYSCALL_64_after_h= wframe > 2.99 =C2=B1 11% -1.2 1.82 =C2=B1 6% perf-profile.cal= ltrace.cycles-pp.path_lookupat.filename_lookup.do_faccessat.do_syscall_64.e= ntry_SYSCALL_64_after_hwframe > 1.55 =C2=B1 11% -1.2 0.39 =C2=B1 57% perf-profile.cal= ltrace.cycles-pp.join_transaction.start_transaction.btrfs_dirty_inode.btrfs= _setattr.notify_change > 2.89 =C2=B1 10% -1.1 1.75 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.vfs_statx.__do_sys_newfstatat.do_syscall_64.entry_SYSCALL_= 64_after_hwframe > 2.46 =C2=B1 7% -1.1 1.33 =C2=B1 2% perf-profile.cal= ltrace.cycles-pp.__btrfs_release_delayed_node.btrfs_delayed_update_inode.bt= rfs_update_inode.btrfs_dirty_inode.btrfs_setattr > 1.79 =C2=B1 11% -0.7 1.11 =C2=B1 6% perf-profile.cal= ltrace.cycles-pp.link_path_walk.path_lookupat.filename_lookup.do_faccessat.= do_syscall_64 > 1.57 =C2=B1 11% -0.7 0.91 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.kmem_cache_alloc.prepare_creds.do_faccessat.do_syscall_64.= entry_SYSCALL_64_after_hwframe > 1.64 =C2=B1 9% -0.6 1.00 =C2=B1 4% perf-profile.cal= ltrace.cycles-pp.user_path_at_empty.do_faccessat.do_syscall_64.entry_SYSCAL= L_64_after_hwframe > 1.60 =C2=B1 9% -0.6 0.98 =C2=B1 4% perf-profile.cal= ltrace.cycles-pp.getname_flags.user_path_at_empty.do_faccessat.do_syscall_6= 4.entry_SYSCALL_64_after_hwframe > 1.54 =C2=B1 8% -0.6 0.94 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.__kmalloc.security_prepare_creds.prepare_creds.do_faccessa= t.do_syscall_64 > 1.21 =C2=B1 10% -0.4 0.77 =C2=B1 5% perf-profile.cal= ltrace.cycles-pp.strncpy_from_user.getname_flags.user_path_at_empty.do_facc= essat.do_syscall_64 > 1.00 =C2=B1 10% -0.4 0.63 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.filename_lookup.vfs_statx.__do_sys_newfstatat.do_syscall_6= 4.entry_SYSCALL_64_after_hwframe > 0.95 =C2=B1 14% -0.4 0.58 =C2=B1 4% perf-profile.cal= ltrace.cycles-pp.kmem_cache_free.do_faccessat.do_syscall_64.entry_SYSCALL_6= 4_after_hwframe > 0.95 =C2=B1 7% -0.3 0.61 =C2=B1 4% perf-profile.cal= ltrace.cycles-pp.kfree.security_cred_free.put_cred_rcu.do_faccessat.do_sysc= all_64 > 0.91 =C2=B1 10% -0.3 0.57 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.path_lookupat.filename_lookup.vfs_statx.__do_sys_newfstata= t.do_syscall_64 > 0.00 +1.0 0.98 =C2=B1 4% perf-profile.calltrac= e.cycles-pp._find_next_bit.cpumask_next.__percpu_counter_sum.__reserve_byte= s.btrfs_reserve_metadata_bytes > 0.00 +1.6 1.56 =C2=B1 4% perf-profile.calltrac= e.cycles-pp.cpumask_next.__percpu_counter_sum.__reserve_bytes.btrfs_reserve= _metadata_bytes.btrfs_block_rsv_add > 0.00 +4.2 4.23 =C2=B1 4% perf-profile.calltrac= e.cycles-pp.__percpu_counter_sum.__reserve_bytes.btrfs_reserve_metadata_byt= es.btrfs_block_rsv_add.start_transaction > 0.00 +12.6 12.55 =C2=B1 4% perf-profile.calltrac= e.cycles-pp.native_queued_spin_lock_slowpath._raw_spin_lock.__reserve_bytes= .btrfs_reserve_metadata_bytes.btrfs_block_rsv_add > 0.00 +13.0 13.01 =C2=B1 4% perf-profile.calltrac= e.cycles-pp._raw_spin_lock.__reserve_bytes.btrfs_reserve_metadata_bytes.btr= fs_block_rsv_add.start_transaction > 0.00 +15.3 15.29 =C2=B1 3% perf-profile.calltrac= e.cycles-pp.native_queued_spin_lock_slowpath._raw_spin_lock.btrfs_block_rsv= _release.btrfs_trans_release_metadata.__btrfs_end_transaction > 0.00 +15.8 15.81 =C2=B1 3% perf-profile.calltrac= e.cycles-pp._raw_spin_lock.btrfs_block_rsv_release.btrfs_trans_release_meta= data.__btrfs_end_transaction.btrfs_dirty_inode > 0.00 +16.0 16.04 =C2=B1 3% perf-profile.calltrac= e.cycles-pp.btrfs_block_rsv_release.btrfs_trans_release_metadata.__btrfs_en= d_transaction.btrfs_dirty_inode.btrfs_setattr > 0.00 +16.0 16.05 =C2=B1 3% perf-profile.calltrac= e.cycles-pp.btrfs_trans_release_metadata.__btrfs_end_transaction.btrfs_dirt= y_inode.btrfs_setattr.notify_change > 0.63 =C2=B1 10% +16.7 17.34 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.__btrfs_end_transaction.btrfs_dirty_inode.btrfs_setattr.no= tify_change.chmod_common > 0.00 +18.3 18.27 =C2=B1 4% perf-profile.calltrac= e.cycles-pp.__reserve_bytes.btrfs_reserve_metadata_bytes.btrfs_block_rsv_ad= d.start_transaction.btrfs_dirty_inode > 0.00 +18.4 18.39 =C2=B1 4% perf-profile.calltrac= e.cycles-pp.btrfs_reserve_metadata_bytes.btrfs_block_rsv_add.start_transact= ion.btrfs_dirty_inode.btrfs_setattr > 2.18 =C2=B1 12% +18.4 20.61 =C2=B1 4% perf-profile.cal= ltrace.cycles-pp.start_transaction.btrfs_dirty_inode.btrfs_setattr.notify_c= hange.chmod_common > 0.00 +18.9 18.88 =C2=B1 4% perf-profile.calltrac= e.cycles-pp.btrfs_block_rsv_add.start_transaction.btrfs_dirty_inode.btrfs_s= etattr.notify_change > 9.24 =C2=B1 10% +32.2 41.40 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.__x64_sys_fchmod.do_syscall_64.entry_SYSCALL_64_after_hwfr= ame > 9.16 =C2=B1 10% +32.2 41.36 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.chmod_common.__x64_sys_fchmod.do_syscall_64.entry_SYSCALL_= 64_after_hwframe > 8.30 =C2=B1 9% +32.6 40.94 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.notify_change.chmod_common.__x64_sys_fchmod.do_syscall_64.= entry_SYSCALL_64_after_hwframe > 8.09 =C2=B1 10% +32.7 40.80 =C2=B1 3% perf-profile.cal= ltrace.cycles-pp.btrfs_setattr.notify_change.chmod_common.__x64_sys_fchmod.= do_syscall_64 > 7.74 =C2=B1 9% +32.8 40.56 =C2=B1 4% perf-profile.cal= ltrace.cycles-pp.btrfs_dirty_inode.btrfs_setattr.notify_change.chmod_common= .__x64_sys_fchmod > 47.72 =C2=B1 11% -27.3 20.41 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.do_faccessat > 20.57 =C2=B1 11% -12.1 8.51 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.prepare_creds > 18.81 =C2=B1 11% -11.9 6.93 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.put_cred_rcu > 5.42 =C2=B1 12% -3.8 1.60 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.free_uid > 5.39 =C2=B1 12% -3.8 1.58 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.refcount_dec_and_lock_irqsave > 5.36 =C2=B1 12% -3.8 1.56 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.refcount_dec_not_one > 5.34 =C2=B1 11% -3.1 2.21 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.key_put > 5.18 =C2=B1 11% -2.4 2.75 =C2=B1 4% perf-profile.chi= ldren.cycles-pp.security_prepare_creds > 4.77 =C2=B1 8% -2.3 2.50 =C2=B1 4% perf-profile.chi= ldren.cycles-pp.btrfs_update_inode > 3.80 =C2=B1 7% -2.0 1.84 =C2=B1 4% perf-profile.chi= ldren.cycles-pp.btrfs_delayed_update_inode > 3.54 =C2=B1 12% -1.8 1.74 =C2=B1 4% perf-profile.chi= ldren.cycles-pp.apparmor_cred_prepare > 3.49 =C2=B1 9% -1.8 1.72 =C2=B1 6% perf-profile.chi= ldren.cycles-pp.security_cred_free > 4.39 =C2=B1 11% -1.7 2.68 =C2=B1 5% perf-profile.chi= ldren.cycles-pp.filename_lookup > 3.93 =C2=B1 11% -1.5 2.40 =C2=B1 5% perf-profile.chi= ldren.cycles-pp.path_lookupat > 2.50 =C2=B1 9% -1.4 1.09 =C2=B1 9% perf-profile.chi= ldren.cycles-pp.apparmor_cred_free > 3.26 =C2=B1 11% -1.3 1.96 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.__do_sys_newfstatat > 2.90 =C2=B1 10% -1.1 1.75 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.vfs_statx > 2.47 =C2=B1 7% -1.1 1.34 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.__btrfs_release_delayed_node > 1.55 =C2=B1 11% -1.1 0.49 =C2=B1 7% perf-profile.chi= ldren.cycles-pp.join_transaction > 2.37 =C2=B1 11% -0.9 1.49 =C2=B1 6% perf-profile.chi= ldren.cycles-pp.link_path_walk > 2.11 =C2=B1 10% -0.9 1.23 =C2=B1 4% perf-profile.chi= ldren.cycles-pp.kmem_cache_alloc > 2.13 =C2=B1 9% -0.8 1.30 =C2=B1 4% perf-profile.chi= ldren.cycles-pp.user_path_at_empty > 2.08 =C2=B1 9% -0.8 1.29 =C2=B1 4% perf-profile.chi= ldren.cycles-pp.getname_flags > 1.17 =C2=B1 12% -0.7 0.50 =C2=B1 6% perf-profile.chi= ldren.cycles-pp.common_perm_cond > 1.58 =C2=B1 14% -0.7 0.92 =C2=B1 7% perf-profile.chi= ldren.cycles-pp.inode_permission > 1.10 =C2=B1 12% -0.6 0.46 =C2=B1 6% perf-profile.chi= ldren.cycles-pp.common_perm > 1.58 =C2=B1 9% -0.6 0.97 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.__kmalloc > 1.56 =C2=B1 11% -0.6 0.99 =C2=B1 4% perf-profile.chi= ldren.cycles-pp.strncpy_from_user > 1.29 =C2=B1 14% -0.5 0.79 =C2=B1 5% perf-profile.chi= ldren.cycles-pp.kmem_cache_free > 1.10 =C2=B1 13% -0.5 0.62 =C2=B1 13% perf-profile.chi= ldren.cycles-pp.mutex_lock > 0.59 =C2=B1 10% -0.4 0.14 =C2=B1 16% perf-profile.chi= ldren.cycles-pp.btrfs_get_or_create_delayed_node > 0.57 =C2=B1 10% -0.4 0.14 =C2=B1 18% perf-profile.chi= ldren.cycles-pp.btrfs_get_delayed_node > 0.83 =C2=B1 13% -0.4 0.46 =C2=B1 4% perf-profile.chi= ldren.cycles-pp.generic_permission > 0.91 =C2=B1 10% -0.4 0.56 =C2=B1 6% perf-profile.chi= ldren.cycles-pp.walk_component > 0.96 =C2=B1 7% -0.4 0.61 =C2=B1 4% perf-profile.chi= ldren.cycles-pp.kfree > 0.60 =C2=B1 14% -0.3 0.26 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.security_path_chmod > 0.65 =C2=B1 10% -0.3 0.31 =C2=B1 7% perf-profile.chi= ldren.cycles-pp.vfs_getattr > 0.64 =C2=B1 10% -0.3 0.30 =C2=B1 7% perf-profile.chi= ldren.cycles-pp.security_inode_getattr > 0.77 =C2=B1 12% -0.3 0.46 =C2=B1 6% perf-profile.chi= ldren.cycles-pp.btrfs_update_root_times > 0.69 =C2=B1 5% -0.3 0.39 =C2=B1 5% perf-profile.chi= ldren.cycles-pp.mutex_unlock > 0.65 =C2=B1 12% -0.3 0.39 =C2=B1 9% perf-profile.chi= ldren.cycles-pp.syscall_return_via_sysret > 0.70 =C2=B1 9% -0.3 0.44 =C2=B1 4% perf-profile.chi= ldren.cycles-pp.__check_object_size > 0.62 =C2=B1 12% -0.3 0.37 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.complete_walk > 0.51 =C2=B1 12% -0.3 0.26 =C2=B1 5% perf-profile.chi= ldren.cycles-pp.get_obj_cgroup_from_current > 0.46 =C2=B1 31% -0.2 0.22 =C2=B1 16% perf-profile.chi= ldren.cycles-pp.revert_creds > 0.59 =C2=B1 12% -0.2 0.36 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.unlazy_walk > 0.48 =C2=B1 11% -0.2 0.28 =C2=B1 5% perf-profile.chi= ldren.cycles-pp.obj_cgroup_charge > 0.58 =C2=B1 10% -0.2 0.38 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.btrfs_getattr > 0.52 =C2=B1 10% -0.2 0.33 =C2=B1 9% perf-profile.chi= ldren.cycles-pp.lookup_fast > 0.47 =C2=B1 20% -0.2 0.29 =C2=B1 15% perf-profile.chi= ldren.cycles-pp.btrfs_permission > 0.33 =C2=B1 12% -0.2 0.15 =C2=B1 8% perf-profile.chi= ldren.cycles-pp.capable_wrt_inode_uidgid > 0.46 =C2=B1 12% -0.2 0.29 perf-profile.children= .cycles-pp.__legitimize_path > 0.32 =C2=B1 12% -0.2 0.14 =C2=B1 10% perf-profile.chi= ldren.cycles-pp.security_capable > 0.31 =C2=B1 12% -0.2 0.13 =C2=B1 11% perf-profile.chi= ldren.cycles-pp.apparmor_capable > 0.32 =C2=B1 12% -0.2 0.15 =C2=B1 8% perf-profile.chi= ldren.cycles-pp.ns_capable_common > 0.45 =C2=B1 9% -0.2 0.28 =C2=B1 10% perf-profile.chi= ldren.cycles-pp.refill_obj_stock > 0.28 =C2=B1 63% -0.2 0.12 =C2=B1 27% perf-profile.chi= ldren.cycles-pp.map_id_up > 0.35 =C2=B1 28% -0.1 0.20 =C2=B1 10% perf-profile.chi= ldren.cycles-pp.cp_new_stat > 0.42 =C2=B1 10% -0.1 0.28 =C2=B1 6% perf-profile.chi= ldren.cycles-pp.__mod_memcg_lruvec_state > 0.36 =C2=B1 10% -0.1 0.23 =C2=B1 8% perf-profile.chi= ldren.cycles-pp.path_put > 0.35 =C2=B1 10% -0.1 0.22 =C2=B1 11% perf-profile.chi= ldren.cycles-pp.dput > 0.35 =C2=B1 9% -0.1 0.22 =C2=B1 9% perf-profile.chi= ldren.cycles-pp.__d_lookup_rcu > 0.32 =C2=B1 10% -0.1 0.19 =C2=B1 12% perf-profile.chi= ldren.cycles-pp.path_init > 0.36 =C2=B1 14% -0.1 0.24 =C2=B1 5% perf-profile.chi= ldren.cycles-pp.__entry_text_start > 0.35 =C2=B1 11% -0.1 0.23 =C2=B1 9% perf-profile.chi= ldren.cycles-pp.___might_sleep > 0.32 =C2=B1 9% -0.1 0.20 =C2=B1 4% perf-profile.chi= ldren.cycles-pp.__check_heap_object > 0.27 =C2=B1 10% -0.1 0.16 =C2=B1 7% perf-profile.chi= ldren.cycles-pp.__list_add_valid > 0.26 =C2=B1 7% -0.1 0.16 =C2=B1 11% perf-profile.chi= ldren.cycles-pp.__might_sleep > 0.23 =C2=B1 15% -0.1 0.13 =C2=B1 5% perf-profile.chi= ldren.cycles-pp._cond_resched > 0.19 =C2=B1 39% -0.1 0.09 =C2=B1 7% perf-profile.chi= ldren.cycles-pp.fill_stack_inode_item > 0.23 =C2=B1 14% -0.1 0.14 =C2=B1 12% perf-profile.chi= ldren.cycles-pp.syscall_enter_from_user_mode > 0.23 =C2=B1 14% -0.1 0.15 =C2=B1 7% perf-profile.chi= ldren.cycles-pp.__legitimize_mnt > 0.18 =C2=B1 14% -0.1 0.10 =C2=B1 11% perf-profile.chi= ldren.cycles-pp.override_creds > 0.17 =C2=B1 15% -0.1 0.10 =C2=B1 11% perf-profile.chi= ldren.cycles-pp.__list_del_entry_valid > 0.18 =C2=B1 13% -0.1 0.10 =C2=B1 12% perf-profile.chi= ldren.cycles-pp.step_into > 0.19 =C2=B1 14% -0.1 0.12 =C2=B1 12% perf-profile.chi= ldren.cycles-pp.lockref_get_not_dead > 0.20 =C2=B1 10% -0.1 0.13 =C2=B1 12% perf-profile.chi= ldren.cycles-pp.__mod_memcg_state > 0.19 =C2=B1 16% -0.1 0.12 =C2=B1 11% perf-profile.chi= ldren.cycles-pp.rcu_read_unlock_strict > 0.15 =C2=B1 14% -0.1 0.09 =C2=B1 13% perf-profile.chi= ldren.cycles-pp.lockref_put_or_lock > 0.17 =C2=B1 11% -0.1 0.12 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.memset_erms > 0.17 =C2=B1 17% -0.1 0.12 =C2=B1 7% perf-profile.chi= ldren.cycles-pp.syscall_exit_to_user_mode > 0.16 =C2=B1 7% -0.1 0.11 =C2=B1 4% perf-profile.chi= ldren.cycles-pp._copy_to_user > 0.10 =C2=B1 21% -0.1 0.04 =C2=B1 58% perf-profile.chi= ldren.cycles-pp.mnt_want_write > 0.11 =C2=B1 9% -0.0 0.07 =C2=B1 13% perf-profile.chi= ldren.cycles-pp.inode_get_bytes > 0.11 =C2=B1 18% -0.0 0.07 =C2=B1 13% perf-profile.chi= ldren.cycles-pp.rcu_all_qs > 0.11 =C2=B1 16% -0.0 0.06 =C2=B1 11% perf-profile.chi= ldren.cycles-pp.security_inode_permission > 0.10 -0.0 0.05 =C2=B1 8% perf-profile.children= .cycles-pp.__fget_light > 0.11 =C2=B1 13% -0.0 0.07 =C2=B1 7% perf-profile.chi= ldren.cycles-pp.entry_SYSCALL_64_safe_stack > 0.12 =C2=B1 12% -0.0 0.08 =C2=B1 8% perf-profile.chi= ldren.cycles-pp.copy_user_enhanced_fast_string > 0.10 =C2=B1 19% -0.0 0.07 =C2=B1 7% perf-profile.chi= ldren.cycles-pp.__put_cred > 0.11 =C2=B1 14% -0.0 0.07 =C2=B1 14% perf-profile.chi= ldren.cycles-pp.btrfs_balance_delayed_items > 0.09 =C2=B1 14% -0.0 0.06 =C2=B1 14% perf-profile.chi= ldren.cycles-pp.exit_to_user_mode_prepare > 0.10 =C2=B1 14% -0.0 0.08 =C2=B1 6% perf-profile.chi= ldren.cycles-pp.__x86_indirect_thunk_rax > 0.08 =C2=B1 10% -0.0 0.06 =C2=B1 15% perf-profile.chi= ldren.cycles-pp.mntput_no_expire > 0.21 =C2=B1 6% +0.1 0.28 =C2=B1 12% perf-profile.chi= ldren.cycles-pp.btrfs_put_transaction > 0.05 =C2=B1 61% +0.1 0.15 =C2=B1 9% perf-profile.chi= ldren.cycles-pp._raw_spin_unlock_irqrestore > 0.00 +0.2 0.21 =C2=B1 10% perf-profile.children= .cycles-pp.find_next_bit > 0.06 =C2=B1 17% +0.3 0.38 =C2=B1 4% perf-profile.chi= ldren.cycles-pp._raw_spin_lock_irqsave > 0.00 +0.4 0.43 =C2=B1 14% perf-profile.children= .cycles-pp.btrfs_get_alloc_profile > 0.00 +0.5 0.46 =C2=B1 14% perf-profile.children= .cycles-pp.calc_available_free_space > 0.00 +0.5 0.48 =C2=B1 2% perf-profile.children= .cycles-pp.btrfs_block_rsv_add_bytes > 0.00 +0.5 0.48 =C2=B1 8% perf-profile.children= .cycles-pp.wait_current_trans > 0.00 +1.0 1.00 =C2=B1 4% perf-profile.children= .cycles-pp._find_next_bit > 0.00 +1.8 1.77 =C2=B1 4% perf-profile.children= .cycles-pp.cpumask_next > 0.00 +4.2 4.24 =C2=B1 4% perf-profile.children= .cycles-pp.__percpu_counter_sum > 0.00 +16.0 16.04 =C2=B1 3% perf-profile.children= .cycles-pp.btrfs_block_rsv_release > 0.00 +16.0 16.05 =C2=B1 3% perf-profile.children= .cycles-pp.btrfs_trans_release_metadata > 0.64 =C2=B1 9% +16.7 17.35 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.__btrfs_end_transaction > 0.00 +18.3 18.27 =C2=B1 4% perf-profile.children= .cycles-pp.__reserve_bytes > 0.00 +18.4 18.39 =C2=B1 4% perf-profile.children= .cycles-pp.btrfs_reserve_metadata_bytes > 2.19 =C2=B1 12% +18.4 20.61 =C2=B1 4% perf-profile.chi= ldren.cycles-pp.start_transaction > 0.00 +18.9 18.88 =C2=B1 4% perf-profile.children= .cycles-pp.btrfs_block_rsv_add > 0.75 =C2=B1 8% +27.1 27.88 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.native_queued_spin_lock_slowpath > 2.32 =C2=B1 10% +28.3 30.66 =C2=B1 3% perf-profile.chi= ldren.cycles-pp._raw_spin_lock > 9.25 =C2=B1 10% +32.2 41.41 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.__x64_sys_fchmod > 9.16 =C2=B1 10% +32.2 41.36 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.chmod_common > 8.31 =C2=B1 9% +32.6 40.95 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.notify_change > 8.10 =C2=B1 10% +32.7 40.80 =C2=B1 3% perf-profile.chi= ldren.cycles-pp.btrfs_setattr > 7.75 =C2=B1 9% +32.8 40.56 =C2=B1 4% perf-profile.chi= ldren.cycles-pp.btrfs_dirty_inode > 13.73 =C2=B1 11% -8.9 4.81 =C2=B1 4% perf-profile.sel= f.cycles-pp.prepare_creds > 5.33 =C2=B1 12% -3.8 1.55 =C2=B1 3% perf-profile.sel= f.cycles-pp.refcount_dec_not_one > 4.57 =C2=B1 11% -3.1 1.42 =C2=B1 4% perf-profile.sel= f.cycles-pp.put_cred_rcu > 5.30 =C2=B1 11% -3.1 2.19 =C2=B1 3% perf-profile.sel= f.cycles-pp.key_put > 3.53 =C2=B1 12% -1.8 1.73 =C2=B1 4% perf-profile.sel= f.cycles-pp.apparmor_cred_prepare > 2.48 =C2=B1 9% -1.4 1.08 =C2=B1 9% perf-profile.sel= f.cycles-pp.apparmor_cred_free > 1.10 =C2=B1 12% -0.6 0.45 =C2=B1 5% perf-profile.sel= f.cycles-pp.common_perm > 1.04 =C2=B1 14% -0.5 0.58 =C2=B1 13% perf-profile.sel= f.cycles-pp.mutex_lock > 0.57 =C2=B1 11% -0.4 0.13 =C2=B1 20% perf-profile.sel= f.cycles-pp.btrfs_get_delayed_node > 1.06 =C2=B1 9% -0.4 0.63 =C2=B1 5% perf-profile.sel= f.cycles-pp.kmem_cache_alloc > 1.02 =C2=B1 10% -0.4 0.64 =C2=B1 4% perf-profile.sel= f.cycles-pp.link_path_walk > 0.93 =C2=B1 14% -0.4 0.55 =C2=B1 5% perf-profile.sel= f.cycles-pp.kmem_cache_free > 0.79 =C2=B1 8% -0.3 0.48 =C2=B1 4% perf-profile.sel= f.cycles-pp.__kmalloc > 0.69 =C2=B1 5% -0.3 0.39 =C2=B1 6% perf-profile.sel= f.cycles-pp.mutex_unlock > 0.72 =C2=B1 11% -0.3 0.46 =C2=B1 5% perf-profile.sel= f.cycles-pp.strncpy_from_user > 0.65 =C2=B1 12% -0.3 0.39 =C2=B1 9% perf-profile.sel= f.cycles-pp.syscall_return_via_sysret > 0.46 =C2=B1 31% -0.2 0.22 =C2=B1 16% perf-profile.sel= f.cycles-pp.revert_creds > 0.64 =C2=B1 9% -0.2 0.40 =C2=B1 3% perf-profile.sel= f.cycles-pp.kfree > 0.60 =C2=B1 9% -0.2 0.38 =C2=B1 7% perf-profile.sel= f.cycles-pp.join_transaction > 0.41 =C2=B1 11% -0.2 0.22 =C2=B1 5% perf-profile.sel= f.cycles-pp.get_obj_cgroup_from_current > 0.45 =C2=B1 10% -0.2 0.27 =C2=B1 5% perf-profile.sel= f.cycles-pp.obj_cgroup_charge > 0.30 =C2=B1 13% -0.2 0.13 =C2=B1 11% perf-profile.sel= f.cycles-pp.apparmor_capable > 0.44 =C2=B1 20% -0.2 0.28 =C2=B1 16% perf-profile.sel= f.cycles-pp.btrfs_permission > 0.42 =C2=B1 11% -0.2 0.27 =C2=B1 8% perf-profile.sel= f.cycles-pp.generic_permission > 0.42 =C2=B1 8% -0.2 0.27 =C2=B1 10% perf-profile.sel= f.cycles-pp.refill_obj_stock > 0.47 =C2=B1 10% -0.2 0.32 =C2=B1 2% perf-profile.sel= f.cycles-pp.btrfs_getattr > 0.27 =C2=B1 65% -0.2 0.12 =C2=B1 25% perf-profile.sel= f.cycles-pp.map_id_up > 0.36 =C2=B1 14% -0.1 0.23 =C2=B1 4% perf-profile.sel= f.cycles-pp.__entry_text_start > 0.34 =C2=B1 10% -0.1 0.22 =C2=B1 9% perf-profile.sel= f.cycles-pp.__d_lookup_rcu > 0.38 =C2=B1 11% -0.1 0.25 =C2=B1 5% perf-profile.sel= f.cycles-pp.do_faccessat > 0.32 =C2=B1 19% -0.1 0.20 =C2=B1 13% perf-profile.sel= f.cycles-pp.inode_permission > 0.34 =C2=B1 13% -0.1 0.22 =C2=B1 9% perf-profile.sel= f.cycles-pp.___might_sleep > 0.31 =C2=B1 9% -0.1 0.20 =C2=B1 4% perf-profile.sel= f.cycles-pp.__check_heap_object > 0.27 =C2=B1 10% -0.1 0.16 =C2=B1 7% perf-profile.sel= f.cycles-pp.__list_add_valid > 0.28 =C2=B1 11% -0.1 0.17 =C2=B1 11% perf-profile.sel= f.cycles-pp.path_init > 0.25 =C2=B1 8% -0.1 0.15 =C2=B1 10% perf-profile.sel= f.cycles-pp.__might_sleep > 0.26 =C2=B1 8% -0.1 0.16 =C2=B1 7% perf-profile.sel= f.cycles-pp.__check_object_size > 0.23 =C2=B1 16% -0.1 0.13 =C2=B1 13% perf-profile.sel= f.cycles-pp.syscall_enter_from_user_mode > 0.17 =C2=B1 16% -0.1 0.09 =C2=B1 7% perf-profile.sel= f.cycles-pp.override_creds > 0.22 =C2=B1 15% -0.1 0.14 =C2=B1 8% perf-profile.sel= f.cycles-pp.__legitimize_mnt > 0.20 =C2=B1 13% -0.1 0.12 =C2=B1 12% perf-profile.sel= f.cycles-pp.entry_SYSCALL_64_after_hwframe > 0.14 =C2=B1 18% -0.1 0.06 perf-profile.self.cyc= les-pp.btrfs_update_inode > 0.21 =C2=B1 13% -0.1 0.13 =C2=B1 5% perf-profile.sel= f.cycles-pp.walk_component > 0.17 =C2=B1 11% -0.1 0.10 =C2=B1 11% perf-profile.sel= f.cycles-pp.step_into > 0.17 =C2=B1 17% -0.1 0.10 =C2=B1 11% perf-profile.sel= f.cycles-pp.__list_del_entry_valid > 0.19 =C2=B1 14% -0.1 0.12 =C2=B1 12% perf-profile.sel= f.cycles-pp.lockref_get_not_dead > 0.20 =C2=B1 10% -0.1 0.13 =C2=B1 12% perf-profile.sel= f.cycles-pp.__mod_memcg_state > 0.16 =C2=B1 10% -0.1 0.10 =C2=B1 11% perf-profile.sel= f.cycles-pp.lookup_fast > 0.15 =C2=B1 14% -0.1 0.08 =C2=B1 13% perf-profile.sel= f.cycles-pp.lockref_put_or_lock > 0.20 =C2=B1 10% -0.1 0.14 =C2=B1 5% perf-profile.sel= f.cycles-pp.__mod_memcg_lruvec_state > 0.15 =C2=B1 14% -0.1 0.08 =C2=B1 10% perf-profile.sel= f.cycles-pp.filename_lookup > 0.17 =C2=B1 9% -0.1 0.11 =C2=B1 3% perf-profile.sel= f.cycles-pp.memset_erms > 0.14 =C2=B1 12% -0.1 0.08 =C2=B1 5% perf-profile.sel= f.cycles-pp.getname_flags > 0.07 =C2=B1 11% -0.0 0.03 =C2=B1100% perf-profile.sel= f.cycles-pp.dput > 0.11 =C2=B1 16% -0.0 0.06 =C2=B1 11% perf-profile.sel= f.cycles-pp.security_inode_permission > 0.15 =C2=B1 15% -0.0 0.10 =C2=B1 14% perf-profile.sel= f.cycles-pp.rcu_read_unlock_strict > 0.12 =C2=B1 12% -0.0 0.08 =C2=B1 10% perf-profile.sel= f.cycles-pp.copy_user_enhanced_fast_string > 0.11 =C2=B1 11% -0.0 0.06 =C2=B1 13% perf-profile.sel= f.cycles-pp._cond_resched > 0.11 =C2=B1 13% -0.0 0.07 =C2=B1 7% perf-profile.sel= f.cycles-pp.entry_SYSCALL_64_safe_stack > 0.10 =C2=B1 5% -0.0 0.05 =C2=B1 8% perf-profile.sel= f.cycles-pp.__fget_light > 0.10 =C2=B1 19% -0.0 0.07 =C2=B1 13% perf-profile.sel= f.cycles-pp.path_lookupat > 0.11 =C2=B1 14% -0.0 0.07 =C2=B1 20% perf-profile.sel= f.cycles-pp.btrfs_balance_delayed_items > 0.10 =C2=B1 14% -0.0 0.07 =C2=B1 10% perf-profile.sel= f.cycles-pp.__virt_addr_valid > 0.10 =C2=B1 21% -0.0 0.06 =C2=B1 6% perf-profile.sel= f.cycles-pp.__put_cred > 0.07 =C2=B1 11% -0.0 0.04 =C2=B1 60% perf-profile.sel= f.cycles-pp.exit_to_user_mode_prepare > 0.09 =C2=B1 20% -0.0 0.06 =C2=B1 17% perf-profile.sel= f.cycles-pp.notify_change > 0.07 =C2=B1 12% -0.0 0.05 =C2=B1 8% perf-profile.sel= f.cycles-pp.mntput_no_expire > 0.21 =C2=B1 6% +0.1 0.28 =C2=B1 12% perf-profile.sel= f.cycles-pp.btrfs_put_transaction > 0.05 =C2=B1 62% +0.1 0.13 =C2=B1 11% perf-profile.sel= f.cycles-pp._raw_spin_unlock_irqrestore > 0.00 +0.1 0.12 =C2=B1 5% perf-profile.self.cyc= les-pp.btrfs_reserve_metadata_bytes > 0.00 +0.1 0.14 =C2=B1 13% perf-profile.self.cyc= les-pp.btrfs_block_rsv_release > 0.00 +0.2 0.16 =C2=B1 18% perf-profile.self.cyc= les-pp.btrfs_get_alloc_profile > 0.00 +0.2 0.18 =C2=B1 7% perf-profile.self.cyc= les-pp.wait_current_trans > 0.46 =C2=B1 15% +0.2 0.64 =C2=B1 6% perf-profile.sel= f.cycles-pp.start_transaction > 0.00 +0.2 0.21 =C2=B1 10% perf-profile.self.cyc= les-pp.find_next_bit > 0.06 =C2=B1 20% +0.3 0.38 =C2=B1 4% perf-profile.sel= f.cycles-pp._raw_spin_lock_irqsave > 0.00 +0.6 0.57 =C2=B1 14% perf-profile.self.cyc= les-pp.__reserve_bytes > 0.00 +0.6 0.57 =C2=B1 7% perf-profile.self.cyc= les-pp.cpumask_next > 0.36 =C2=B1 12% +0.6 0.98 =C2=B1 11% perf-profile.sel= f.cycles-pp.__btrfs_end_transaction > 0.00 +1.0 0.99 =C2=B1 4% perf-profile.self.cyc= les-pp._find_next_bit > 1.57 =C2=B1 11% +1.2 2.77 =C2=B1 6% perf-profile.sel= f.cycles-pp._raw_spin_lock > 0.00 +2.1 2.08 =C2=B1 4% perf-profile.self.cyc= les-pp.__percpu_counter_sum > 0.73 =C2=B1 8% +27.1 27.79 =C2=B1 3% perf-profile.sel= f.cycles-pp.native_queued_spin_lock_slowpath > > > > stress-ng.access.ops > > 1.2e+06 +-------------------------------------------------------------= ----+ > | +..+..+..+.. .+..+..+. .+.. .+..+..+..+..+..+.= .+ | > 1e+06 |-+ : +. +. +. = | > | : = | > | : = | > 800000 |-+ : = | > | : O O O O O O O O O O O = O | > 600000 |-+O :O O O O O O O O = | > | : = | > 400000 |-+ : = | > | : = | > | : = | > 200000 |-+: = | > | : = | > 0 +-------------------------------------------------------------= ----+ > > > stress-ng.access.ops_per_sec > > 40000 +---------------------------------------------------------------= ----+ > | +..+..+..+... .+..+..+. .+.. .+..+...+..+..+..+.= .+ | > 35000 |-+ : +. +. +. = | > 30000 |-+ : = | > | : = | > 25000 |-+ : = | > | : O O O O O O O O O O O = O | > 20000 |-+O :O O O O O O O O = | > | : = | > 15000 |-+ : = | > 10000 |-+ : = | > | : = | > 5000 |-+: = | > | : = | > 0 +---------------------------------------------------------------= ----+ > > > stress-ng.chmod.ops > > 140000 +--------------------------------------------------------------= ----+ > | +.. = | > 120000 |-+ : +..+.. .+.. = | > | : +.. .. +. +.. +.. +.. +.. = | > 100000 |-+ : + ..+.. .. .. .. +.. = | > | : +. + + + +..+.= .+ | > 80000 |-+ : = | > | : = | > 60000 |-+ : O O O = | > | O: O O O O O O O O O O O O O O O O = O | > 40000 |-+ : = | > | : = | > 20000 |-+: = | > | : = | > 0 +--------------------------------------------------------------= ----+ > > > stress-ng.chmod.ops_per_sec > > 4500 +----------------------------------------------------------------= ----+ > | +.. +.. .+.. = | > 4000 |-+ : + +..+. +.. +.. = | > 3500 |-+ : +... + .+.. .+.. + = | > | : + +...+..+. +. + +... .+.= .+ | > 3000 |-+ : + +. = | > 2500 |-+ : = | > | : = | > 2000 |-+ : O O O O O O O = | > 1500 |-+O: O O O O O O O O O O O O = O | > | : = | > 1000 |-+ : = | > 500 |-+: = | > | : = | > 0 +----------------------------------------------------------------= ----+ > > > stress-ng.chown.ops > > 1.2e+06 +-------------------------------------------------------------= ----+ > | +..+..+.. .+.. +.. = | > 1e+06 |-+ : +. +.. .+.. .. .+..+.. .+.. = | > | : +. + +..+..+..+. +. = .+ | > | : +.= | > 800000 |-+ : = | > | : = | > 600000 |-+O :O O O O O O O O = | > | : O O O O O O O O O O O = O | > 400000 |-+ : = | > | : = | > | : = | > 200000 |-+: = | > | : = | > 0 +-------------------------------------------------------------= ----+ > > > stress-ng.chown.ops_per_sec > > 40000 +---------------------------------------------------------------= ----+ > | +..+..+.. ..+.. +.. = | > 35000 |-+ : +. +.. .+.. .. .+...+.. .+.. = | > 30000 |-+ : +. + +..+..+..+. +. = .+ | > | : +.= | > 25000 |-+ : = | > | : = | > 20000 |-+O :O O O O O O O O = | > | : O O O O O O O O O O O = O | > 15000 |-+ : = | > 10000 |-+ : = | > | : = | > 5000 |-+: = | > | : = | > 0 +---------------------------------------------------------------= ----+ > > > stress-ng.utime.ops > > 4.5e+06 +-------------------------------------------------------------= ----+ > | .+..+.. .+.. .+.. .+.. = | > 4e+06 |-+ +..+. +. +. +. .+..+..+..+.. = .+ | > 3.5e+06 |-+ : +..+..+..+. +.= | > | : = | > 3e+06 |-+ : = | > 2.5e+06 |-+ : = | > | : = | > 2e+06 |-+ : = | > 1.5e+06 |-+ : = | > | : O O O O O = | > 1e+06 |-+O: O O O O O O O O O O O O O O = O | > 500000 |-+ : = | > | : = | > 0 +-------------------------------------------------------------= ----+ > > > stress-ng.utime.ops_per_sec > > 160000 +--------------------------------------------------------------= ----+ > | .+.. +.. = | > 140000 |-+ .+. +.. .. .+.. ..+.. = | > 120000 |-+ +. + +. +. .+.. .+..+..+..+..+.. = .+ | > | : +. +. +.= | > 100000 |-+ : = | > | : = | > 80000 |-+ : = | > | : = | > 60000 |-+ : = | > 40000 |-+ : = | > | O: O O O O O O O O O O O O O O O O O O O = O | > 20000 |-+ : = | > | : = | > 0 +--------------------------------------------------------------= ----+ > > > [*] bisect-good sample > [O] bisect-bad sample > > > > Disclaimer: > Results have been estimated based on internal Intel analysis and are prov= ided > for informational purposes only. Any difference in system hardware or sof= tware > design or configuration may affect actual performance. > > > Thanks, > Oliver Sang > --===============0663821624838511493==--