Greeting, FYI, we noticed a -6.9% regression of fio.write_bw_MBps due to commit: commit: 4fd93529bcf2686433120ac18662513548d937d3 ("btrfs: introduce delayed_refs_rsv") https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master in testcase: fio-basic on test machine: 72 threads Intel(R) Xeon(R) CPU E5-2699 v3 @ 2.30GHz with 128G memory with following parameters: disk: 1SSD fs: btrfs runtime: 300s nr_task: 8 rw: randwrite bs: 4k ioengine: sync test_size: 512g ucode: 0x3d cpufreq_governor: performance test-description: Fio is a tool that will spawn a number of threads or processes doing a particular type of I/O action as specified by the user. test-url: https://github.com/axboe/fio 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 ========================================================================================= bs/compiler/cpufreq_governor/disk/fs/ioengine/kconfig/nr_task/rootfs/runtime/rw/tbox_group/test_size/testcase/ucode: 4k/gcc-7/performance/1SSD/btrfs/sync/x86_64-rhel-7.2/8/debian-x86_64-2018-04-03.cgz/300s/randwrite/lkp-hsw-ep2/512g/fio-basic/0x3d commit: c3b63109d1 ("btrfs: only track ref_heads in delayed_ref_updates") 4fd93529bc ("btrfs: introduce delayed_refs_rsv") c3b63109d175d206 4fd93529bcf2686433120ac186 ---------------- -------------------------- fail:runs %reproduction fail:runs | | | 2:4 -50% :4 dmesg.WARNING:at#for_ip_interrupt_entry/0x %stddev %change %stddev \ | \ 0.19 ± 10% +0.8 0.95 ± 3% fio.latency_1000us% 1.76 ± 8% +21.6 23.40 ± 5% fio.latency_100us% 0.02 ± 3% +0.1 0.09 ± 10% fio.latency_10ms% 5.24 ± 8% -1.6 3.66 ± 11% fio.latency_10us% 0.01 +0.0 0.02 ± 11% fio.latency_20ms% 1.55 ± 3% -0.7 0.86 ± 27% fio.latency_20us% 75.43 -52.2 23.19 ± 2% fio.latency_250us% 0.49 ± 4% +0.5 0.94 ± 9% fio.latency_2ms% 0.02 ± 17% +0.2 0.26 ± 11% fio.latency_4ms% 14.71 ± 7% -3.4 11.33 ± 5% fio.latency_500us% 0.20 ± 9% +32.1 32.33 fio.latency_50us% 0.21 ± 13% +2.6 2.80 ± 4% fio.latency_750us% 325.40 -1.8% 319.61 fio.time.elapsed_time 325.40 -1.8% 319.61 fio.time.elapsed_time.max 71398082 -6.6% 66663998 fio.time.file_system_outputs 24.71 ± 2% -10.6% 22.10 fio.time.user_time 8326551 -4.2% 7980852 fio.time.voluntary_context_switches 8922305 -6.6% 8331632 fio.workload 116.17 -6.9% 108.10 fio.write_bw_MBps 274.00 +28.5% 352.00 fio.write_clat_90%_us 316.00 +59.8% 505.00 fio.write_clat_95%_us 467.00 ± 3% +151.0% 1172 ± 4% fio.write_clat_99%_us 267.51 +7.6% 287.72 fio.write_clat_mean_us 29740 -6.9% 27673 fio.write_iops 9896739 ± 11% +34.2% 13283798 ± 6% cpuidle.POLL.time 3005327 ± 12% +36.3% 4097378 ± 4% cpuidle.POLL.usage 9059 ± 21% -20.1% 7235 ± 2% softirqs.NET_RX 5980740 ± 4% -17.0% 4961204 ± 6% softirqs.RCU 198.25 ± 2% +10.6% 219.25 ± 2% turbostat.Avg_MHz 11.89 ± 10% -30.7% 8.24 ± 24% turbostat.Pkg%pc2 0.36 ± 2% +0.4 0.77 ± 4% mpstat.cpu.iowait% 4.26 +0.7 4.98 ± 2% mpstat.cpu.sys% 0.13 ± 2% -0.0 0.11 ± 9% mpstat.cpu.usr% 573529 ± 10% -39.8% 345193 ± 12% numa-meminfo.node0.Dirty 21728 ± 7% +49.3% 32433 ± 10% numa-meminfo.node0.Writeback 8778 ± 15% +77.0% 15540 ± 13% numa-meminfo.node1.Writeback 143494 ± 10% -40.1% 85936 ± 12% numa-vmstat.node0.nr_dirty 5496 ± 5% +53.1% 8414 ± 6% numa-vmstat.node0.nr_writeback 149300 ± 9% -36.6% 94629 ± 11% numa-vmstat.node0.nr_zone_write_pending 2184 ± 21% +71.9% 3754 ± 14% numa-vmstat.node1.nr_writeback 218022 +10.2% 240277 vmstat.io.bo 26855227 -8.6% 24557716 vmstat.memory.cache 396691 +9.3% 433456 vmstat.system.cs 180728 -1.7% 177599 vmstat.system.in 54.15 ± 2% +69.9% 91.99 ± 4% iostat.sdb.avgqu-sz 202.81 ± 9% -32.3% 137.37 ± 10% iostat.sdb.avgrq-sz 27069 -4.0% 25994 iostat.sdb.w/s 217862 +10.3% 240272 iostat.sdb.wkB/s 1.37 ± 20% +261.6% 4.96 ± 5% iostat.sdb.wrqm/s 5028110 ± 6% -25.2% 3762248 ± 5% meminfo.Active 4751087 ± 6% -26.7% 3483276 ± 6% meminfo.Active(file) 25867858 -8.8% 23583792 meminfo.Cached 887304 ± 2% -31.5% 607410 meminfo.Dirty 30842 ± 8% +54.3% 47603 ± 4% meminfo.Writeback 5.038e+11 ± 5% +14.4% 5.762e+11 ± 5% perf-stat.branch-instructions 1.312e+08 +6.5% 1.398e+08 perf-stat.context-switches 4.936e+12 ± 3% +9.3% 5.395e+12 ± 6% perf-stat.cpu-cycles 0.55 ± 12% -0.1 0.40 ± 8% perf-stat.dTLB-load-miss-rate% 3.947e+09 ± 12% -16.4% 3.302e+09 ± 9% perf-stat.dTLB-load-misses 7.115e+11 ± 6% +14.4% 8.141e+11 ± 5% perf-stat.dTLB-loads 2.637e+12 ± 6% +14.1% 3.009e+12 ± 5% perf-stat.instructions 903201 -2.9% 877236 perf-stat.minor-faults 903221 -2.9% 877223 perf-stat.page-faults 295526 ± 6% +22.2% 361161 ± 5% perf-stat.path-length 233.51 ± 28% +263.3% 848.24 ± 26% sched_debug.cfs_rq:/.exec_clock.min 7673 ± 14% -21.6% 6016 ± 14% sched_debug.cfs_rq:/.exec_clock.stddev 195.54 ± 10% +14.8% 224.58 ± 5% sched_debug.cfs_rq:/.load_avg.stddev 2.73 ± 28% +52.7% 4.18 ± 27% sched_debug.cfs_rq:/.removed.util_avg.avg 13.44 ± 14% +70.9% 22.96 ± 34% sched_debug.cfs_rq:/.removed.util_avg.stddev 41.17 ± 24% +37.7% 56.70 ± 9% sched_debug.cfs_rq:/.runnable_load_avg.avg 138.20 ± 12% +16.3% 160.67 ± 3% sched_debug.cfs_rq:/.runnable_load_avg.stddev 91.34 ± 10% +15.2% 105.27 ± 10% sched_debug.cpu.cpu_load[2].stddev 79.46 ± 10% +18.8% 94.42 ± 11% sched_debug.cpu.cpu_load[3].stddev 814.50 +11.6% 908.82 ± 7% sched_debug.cpu.curr->pid.stddev 12393 ± 56% +458.8% 69261 ± 13% sched_debug.cpu.nr_switches.min 880934 ± 11% -23.2% 676786 ± 14% sched_debug.cpu.nr_switches.stddev 12090 ± 57% +467.5% 68614 ± 13% sched_debug.cpu.sched_count.min 881104 ± 11% -23.1% 677137 ± 14% sched_debug.cpu.sched_count.stddev 6021 ± 57% +469.2% 34275 ± 13% sched_debug.cpu.sched_goidle.min 440105 ± 11% -23.2% 338119 ± 14% sched_debug.cpu.sched_goidle.stddev 1512 ± 64% +688.4% 11921 ± 42% sched_debug.cpu.ttwu_count.min 3574 ± 24% +46.1% 5220 ± 18% sched_debug.cpu.ttwu_local.avg 138459 ± 47% +70.4% 235958 ± 28% sched_debug.cpu.ttwu_local.max 16146 ± 47% +70.3% 27493 ± 28% sched_debug.cpu.ttwu_local.stddev 1187638 ± 6% -26.7% 870813 ± 6% proc-vmstat.nr_active_file 18056001 +7.5% 19407987 proc-vmstat.nr_dirtied 222170 ± 2% -31.7% 151680 proc-vmstat.nr_dirty 6469411 -8.8% 5898305 proc-vmstat.nr_file_pages 25594617 +2.3% 26176799 proc-vmstat.nr_free_pages 55526 +2.8% 57092 proc-vmstat.nr_inactive_anon 4989952 -5.1% 4734307 proc-vmstat.nr_inactive_file 59395 +2.7% 60992 proc-vmstat.nr_mapped 57213 +2.6% 58672 proc-vmstat.nr_shmem 480860 -2.6% 468192 proc-vmstat.nr_slab_unreclaimable 7572 ± 7% +59.8% 12102 ± 4% proc-vmstat.nr_writeback 18027209 +7.5% 19370852 proc-vmstat.nr_written 1187638 ± 6% -26.7% 870813 ± 6% proc-vmstat.nr_zone_active_file 55526 +2.8% 57092 proc-vmstat.nr_zone_inactive_anon 4989952 -5.1% 4734307 proc-vmstat.nr_zone_inactive_file 230210 ± 2% -28.7% 164220 proc-vmstat.nr_zone_write_pending 13092479 -7.9% 12051991 proc-vmstat.numa_hit 13078305 -7.9% 12038637 proc-vmstat.numa_local 1203496 ± 7% -50.0% 601160 ± 12% proc-vmstat.pgactivate 16801820 ± 2% +5.8% 17780491 proc-vmstat.pgalloc_normal 923922 -2.9% 897337 proc-vmstat.pgfault 72108783 +7.5% 77484272 proc-vmstat.pgpgout 29951 ± 9% -36.6% 18992 ± 2% slabinfo.Acpi-ParseExt.active_objs 784.50 ± 9% -34.2% 516.25 ± 2% slabinfo.Acpi-ParseExt.active_slabs 30612 ± 9% -34.2% 20148 ± 2% slabinfo.Acpi-ParseExt.num_objs 784.50 ± 9% -34.2% 516.25 ± 2% slabinfo.Acpi-ParseExt.num_slabs 47360 ± 5% -41.6% 27645 ± 5% slabinfo.avc_xperms_data.active_objs 372.75 ± 5% -41.1% 219.50 ± 4% slabinfo.avc_xperms_data.active_slabs 47782 ± 5% -41.0% 28171 ± 5% slabinfo.avc_xperms_data.num_objs 372.75 ± 5% -41.1% 219.50 ± 4% slabinfo.avc_xperms_data.num_slabs 625.00 ± 24% +51.6% 947.25 ± 7% slabinfo.biovec-128.active_objs 642.75 ± 22% +47.8% 949.75 ± 7% slabinfo.biovec-128.num_objs 753.75 ± 16% +43.5% 1081 ± 5% slabinfo.biovec-64.active_objs 753.75 ± 16% +43.5% 1081 ± 5% slabinfo.biovec-64.num_objs 2199 ± 24% +126.4% 4977 ± 6% slabinfo.biovec-max.active_objs 555.00 ± 24% +125.6% 1252 ± 5% slabinfo.biovec-max.active_slabs 2222 ± 24% +125.4% 5010 ± 5% slabinfo.biovec-max.num_objs 555.00 ± 24% +125.6% 1252 ± 5% slabinfo.biovec-max.num_slabs 33746 ± 8% -37.0% 21253 slabinfo.btrfs_delayed_ref_head.active_objs 674.50 ± 8% -34.8% 439.75 ± 2% slabinfo.btrfs_delayed_ref_head.active_slabs 34430 ± 8% -34.8% 22460 ± 2% slabinfo.btrfs_delayed_ref_head.num_objs 674.50 ± 8% -34.8% 439.75 ± 2% slabinfo.btrfs_delayed_ref_head.num_slabs 725.75 ± 2% -12.8% 632.50 ± 4% slabinfo.btrfs_inode.active_objs 727.75 ± 2% -13.0% 633.25 ± 4% slabinfo.btrfs_inode.num_objs 19890 ± 2% +64.6% 32735 ± 3% slabinfo.btrfs_ordered_extent.active_objs 513.25 +64.7% 845.25 ± 3% slabinfo.btrfs_ordered_extent.active_slabs 20038 +64.6% 32983 ± 3% slabinfo.btrfs_ordered_extent.num_objs 513.25 +64.7% 845.25 ± 3% slabinfo.btrfs_ordered_extent.num_slabs 8562 -37.4% 5357 slabinfo.kmalloc-192.active_objs 8576 -37.4% 5371 slabinfo.kmalloc-192.num_objs 79749 ± 3% +22.0% 97286 slabinfo.kmalloc-64.active_objs 1253 ± 3% +22.2% 1531 slabinfo.kmalloc-64.active_slabs 80232 ± 3% +22.2% 98070 slabinfo.kmalloc-64.num_objs 1253 ± 3% +22.2% 1531 slabinfo.kmalloc-64.num_slabs 9144 ± 14% +22.4% 11193 ± 7% slabinfo.mnt_cache.active_objs 9317 ± 13% +22.6% 11425 ± 7% slabinfo.mnt_cache.num_objs 905.50 ± 11% +13.9% 1031 ± 4% slabinfo.skbuff_fclone_cache.active_objs 905.50 ± 11% +13.9% 1031 ± 4% slabinfo.skbuff_fclone_cache.num_objs 7.88 ± 29% -4.8 3.06 ± 54% perf-profile.calltrace.cycles-pp.wb_workfn.process_one_work.worker_thread.kthread.ret_from_fork 7.88 ± 29% -4.8 3.06 ± 54% perf-profile.calltrace.cycles-pp.wb_writeback.wb_workfn.process_one_work.worker_thread.kthread 6.90 ± 26% -4.2 2.71 ± 49% perf-profile.calltrace.cycles-pp.do_writepages.__writeback_single_inode.writeback_sb_inodes.wb_writeback.wb_workfn 6.90 ± 26% -4.2 2.71 ± 49% perf-profile.calltrace.cycles-pp.writeback_sb_inodes.wb_writeback.wb_workfn.process_one_work.worker_thread 6.90 ± 26% -4.2 2.71 ± 49% perf-profile.calltrace.cycles-pp.__writeback_single_inode.writeback_sb_inodes.wb_writeback.wb_workfn.process_one_work 5.64 ± 25% -2.9 2.78 ± 60% perf-profile.calltrace.cycles-pp.extent_write_cache_pages.extent_writepages.do_writepages.__writeback_single_inode.writeback_sb_inodes 5.38 ± 25% -2.7 2.66 ± 61% perf-profile.calltrace.cycles-pp.__extent_writepage.extent_write_cache_pages.extent_writepages.do_writepages.__writeback_single_inode 4.49 ± 24% -2.4 2.14 ± 66% perf-profile.calltrace.cycles-pp.writepage_delalloc.__extent_writepage.extent_write_cache_pages.extent_writepages.do_writepages 4.76 ± 18% -2.3 2.43 ± 57% perf-profile.calltrace.cycles-pp.extent_writepages.do_writepages.__writeback_single_inode.writeback_sb_inodes.wb_writeback 4.04 ± 28% -2.0 2.02 ± 67% perf-profile.calltrace.cycles-pp.btrfs_run_delalloc_range.writepage_delalloc.__extent_writepage.extent_write_cache_pages.extent_writepages 4.04 ± 28% -2.0 2.01 ± 67% perf-profile.calltrace.cycles-pp.run_delalloc_nocow.btrfs_run_delalloc_range.writepage_delalloc.__extent_writepage.extent_write_cache_pages 1.67 ± 45% -1.2 0.45 ±100% perf-profile.calltrace.cycles-pp.submit_one_bio.submit_extent_page.write_one_eb.btree_write_cache_pages.do_writepages 1.67 ± 45% -1.2 0.45 ±100% perf-profile.calltrace.cycles-pp.btree_submit_bio_hook.submit_one_bio.submit_extent_page.write_one_eb.btree_write_cache_pages 1.66 ± 45% -1.2 0.44 ±100% perf-profile.calltrace.cycles-pp.btree_csum_one_bio.btree_submit_bio_hook.submit_one_bio.submit_extent_page.write_one_eb 1.65 ± 45% -1.2 0.44 ±100% perf-profile.calltrace.cycles-pp.csum_tree_block.btree_csum_one_bio.btree_submit_bio_hook.submit_one_bio.submit_extent_page 1.64 ± 45% -1.2 0.44 ±100% perf-profile.calltrace.cycles-pp.crc32c.csum_tree_block.btree_csum_one_bio.btree_submit_bio_hook.submit_one_bio 1.63 ± 45% -1.2 0.44 ±100% perf-profile.calltrace.cycles-pp.chksum_update.crc32c.csum_tree_block.btree_csum_one_bio.btree_submit_bio_hook 1.63 ± 45% -1.2 0.44 ±100% perf-profile.calltrace.cycles-pp.__crc32c_le.chksum_update.crc32c.csum_tree_block.btree_csum_one_bio 0.69 ± 16% +0.4 1.07 ± 28% perf-profile.calltrace.cycles-pp.push_leaf_right.split_leaf.setup_leaf_for_split.btrfs_duplicate_item.btrfs_mark_extent_written 0.82 ± 17% +0.4 1.22 ± 27% perf-profile.calltrace.cycles-pp.split_leaf.setup_leaf_for_split.btrfs_duplicate_item.btrfs_mark_extent_written.btrfs_finish_ordered_io 0.72 ± 18% +0.5 1.18 ± 25% perf-profile.calltrace.cycles-pp.setup_items_for_insert.btrfs_insert_empty_items.btrfs_csum_file_blocks.add_pending_csums.btrfs_finish_ordered_io 0.79 ± 19% +0.5 1.26 ± 21% perf-profile.calltrace.cycles-pp.setup_items_for_insert.btrfs_duplicate_item.btrfs_mark_extent_written.btrfs_finish_ordered_io.normal_work_helper 0.46 ± 59% +0.6 1.02 ± 28% perf-profile.calltrace.cycles-pp.btrfs_search_slot.btrfs_insert_empty_items.btrfs_csum_file_blocks.add_pending_csums.btrfs_finish_ordered_io 0.90 ± 26% +0.6 1.47 ± 27% perf-profile.calltrace.cycles-pp.save_stack_trace_tsk.__account_scheduler_latency.enqueue_entity.enqueue_task_fair.ttwu_do_activate 0.59 ± 60% +0.6 1.19 ± 37% perf-profile.calltrace.cycles-pp.enqueue_task_fair.ttwu_do_activate.try_to_wake_up.autoremove_wake_function.__wake_up_common 0.61 ± 60% +0.6 1.23 ± 37% perf-profile.calltrace.cycles-pp.ttwu_do_activate.try_to_wake_up.autoremove_wake_function.__wake_up_common.__wake_up_common_lock 0.95 ± 29% +0.7 1.68 ± 36% perf-profile.calltrace.cycles-pp.__wake_up_common.__wake_up_common_lock.unlock_up.btrfs_search_slot.btrfs_mark_extent_written 0.78 ± 59% +0.8 1.59 ± 36% perf-profile.calltrace.cycles-pp.try_to_wake_up.autoremove_wake_function.__wake_up_common.__wake_up_common_lock.unlock_up 0.81 ± 59% +0.8 1.66 ± 36% perf-profile.calltrace.cycles-pp.autoremove_wake_function.__wake_up_common.__wake_up_common_lock.unlock_up.btrfs_search_slot 1.33 ± 17% +0.9 2.21 ± 26% perf-profile.calltrace.cycles-pp.btrfs_insert_empty_items.btrfs_csum_file_blocks.add_pending_csums.btrfs_finish_ordered_io.normal_work_helper 2.00 ± 18% +1.2 3.23 ± 27% perf-profile.calltrace.cycles-pp.add_pending_csums.btrfs_finish_ordered_io.normal_work_helper.process_one_work.worker_thread 2.00 ± 17% +1.2 3.23 ± 27% perf-profile.calltrace.cycles-pp.btrfs_csum_file_blocks.add_pending_csums.btrfs_finish_ordered_io.normal_work_helper.process_one_work 2.94 ± 23% +1.5 4.45 ± 16% perf-profile.calltrace.cycles-pp.btrfs_duplicate_item.btrfs_mark_extent_written.btrfs_finish_ordered_io.normal_work_helper.process_one_work 7.88 ± 29% -4.8 3.06 ± 54% perf-profile.children.cycles-pp.wb_workfn 7.88 ± 29% -4.8 3.06 ± 54% perf-profile.children.cycles-pp.wb_writeback 7.88 ± 29% -4.8 3.06 ± 54% perf-profile.children.cycles-pp.writeback_sb_inodes 7.88 ± 29% -4.8 3.06 ± 54% perf-profile.children.cycles-pp.__writeback_single_inode 7.91 ± 29% -4.4 3.48 ± 59% perf-profile.children.cycles-pp.do_writepages 5.74 ± 23% -3.0 2.79 ± 60% perf-profile.children.cycles-pp.extent_writepages 5.74 ± 23% -3.0 2.79 ± 60% perf-profile.children.cycles-pp.extent_write_cache_pages 5.47 ± 23% -2.8 2.67 ± 61% perf-profile.children.cycles-pp.__extent_writepage 4.56 ± 22% -2.3 2.23 ± 61% perf-profile.children.cycles-pp.writepage_delalloc 4.22 ± 22% -2.1 2.10 ± 62% perf-profile.children.cycles-pp.btrfs_run_delalloc_range 4.21 ± 22% -2.1 2.10 ± 62% perf-profile.children.cycles-pp.run_delalloc_nocow 2.47 ± 39% -1.6 0.90 ± 64% perf-profile.children.cycles-pp.submit_extent_page 2.17 ± 47% -1.5 0.70 ± 69% perf-profile.children.cycles-pp.btree_write_cache_pages 2.23 ± 40% -1.5 0.78 ± 67% perf-profile.children.cycles-pp.submit_one_bio 2.08 ± 47% -1.4 0.66 ± 69% perf-profile.children.cycles-pp.write_one_eb 2.44 ± 29% -1.3 1.10 ± 18% perf-profile.children.cycles-pp.crc32c 2.43 ± 29% -1.3 1.10 ± 18% perf-profile.children.cycles-pp.chksum_update 2.43 ± 29% -1.3 1.10 ± 18% perf-profile.children.cycles-pp.__crc32c_le 1.69 ± 46% -1.2 0.54 ± 69% perf-profile.children.cycles-pp.btree_submit_bio_hook 1.68 ± 46% -1.1 0.54 ± 69% perf-profile.children.cycles-pp.btree_csum_one_bio 1.67 ± 46% -1.1 0.54 ± 69% perf-profile.children.cycles-pp.csum_tree_block 1.14 ± 21% -0.5 0.65 ± 30% perf-profile.children.cycles-pp.queue_work_on 0.88 ± 25% -0.5 0.42 ± 62% perf-profile.children.cycles-pp.__extent_writepage_io 1.07 ± 17% -0.4 0.62 ± 29% perf-profile.children.cycles-pp.__queue_work 0.76 ± 21% -0.4 0.35 ± 59% perf-profile.children.cycles-pp.create_io_em 0.60 ± 27% -0.4 0.23 ± 51% perf-profile.children.cycles-pp.btrfs_lookup_csums_range 0.60 ± 27% -0.4 0.23 ± 51% perf-profile.children.cycles-pp.csum_exist_in_range 0.65 ± 25% -0.3 0.31 ± 57% perf-profile.children.cycles-pp.btrfs_drop_extent_cache 0.53 ± 23% -0.3 0.23 ± 71% perf-profile.children.cycles-pp.btrfs_submit_bio_hook 0.53 ± 23% -0.3 0.23 ± 72% perf-profile.children.cycles-pp.btrfs_wq_submit_bio 0.55 ± 27% -0.3 0.28 ± 38% perf-profile.children.cycles-pp.btrfs_inode_rsv_release 0.38 ± 31% -0.2 0.13 ± 76% perf-profile.children.cycles-pp.btrfs_cross_ref_exist 0.33 ± 26% -0.2 0.12 ± 70% perf-profile.children.cycles-pp.find_lock_delalloc_range 0.56 ± 26% -0.2 0.36 ± 40% perf-profile.children.cycles-pp.btrfs_remove_ordered_extent 0.26 ± 27% -0.2 0.10 ± 69% perf-profile.children.cycles-pp.find_delalloc_range 0.22 ± 34% -0.1 0.08 ± 81% perf-profile.children.cycles-pp.check_committed_ref 0.38 ± 13% -0.1 0.25 ± 33% perf-profile.children.cycles-pp.btrfs_map_bio 0.21 ± 34% -0.1 0.09 ± 87% perf-profile.children.cycles-pp.__btrfs_add_ordered_extent 0.21 ± 34% -0.1 0.09 ± 87% perf-profile.children.cycles-pp.btrfs_add_ordered_extent 0.22 ± 13% -0.1 0.11 ± 48% perf-profile.children.cycles-pp.btrfs_tree_read_unlock_blocking 0.22 ± 24% -0.1 0.10 ± 25% perf-profile.children.cycles-pp.btrfs_tree_read_lock_atomic 0.17 ± 28% -0.1 0.05 ±100% perf-profile.children.cycles-pp.end_bio_extent_buffer_writepage 0.17 ± 24% -0.1 0.06 ± 65% perf-profile.children.cycles-pp.extent_clear_unlock_delalloc 0.36 ± 15% -0.1 0.25 ± 35% perf-profile.children.cycles-pp.btrfs_submit_bio_done 0.17 ± 32% -0.1 0.06 ± 65% perf-profile.children.cycles-pp.find_get_pages_range_tag 0.17 ± 32% -0.1 0.06 ± 67% perf-profile.children.cycles-pp.pagevec_lookup_range_tag 0.13 ± 46% -0.1 0.03 ±102% perf-profile.children.cycles-pp.memzero_extent_buffer 0.12 ± 42% -0.1 0.03 ±100% perf-profile.children.cycles-pp.clear_page_dirty_for_io 0.13 ± 34% -0.1 0.04 ±110% perf-profile.children.cycles-pp.__test_set_page_writeback 0.16 ± 25% -0.1 0.08 ± 58% perf-profile.children.cycles-pp.xas_find_marked 0.10 ± 36% -0.1 0.03 ±102% perf-profile.children.cycles-pp.btrfs_merge_bio_hook 0.10 ± 34% -0.1 0.03 ±102% perf-profile.children.cycles-pp.btrfs_map_block 0.11 ± 24% -0.1 0.05 ± 62% perf-profile.children.cycles-pp.btrfs_get_chunk_map 0.11 ± 14% -0.0 0.08 ± 23% perf-profile.children.cycles-pp.leave_mm 0.11 ± 4% +0.0 0.14 ± 8% perf-profile.children.cycles-pp.rb_next 0.08 ± 6% +0.0 0.10 ± 10% perf-profile.children.cycles-pp.tick_nohz_idle_enter 0.23 ± 4% +0.0 0.26 ± 5% perf-profile.children.cycles-pp.dequeue_entity 0.07 ± 17% +0.0 0.10 ± 4% perf-profile.children.cycles-pp.rb_erase 0.16 ± 6% +0.0 0.20 ± 10% perf-profile.children.cycles-pp.nr_iowait_cpu 0.14 ± 11% +0.0 0.18 ± 8% perf-profile.children.cycles-pp.__module_text_address 0.00 +0.1 0.05 ± 8% perf-profile.children.cycles-pp.btrfs_set_token_64 0.28 ± 2% +0.1 0.34 ± 7% perf-profile.children.cycles-pp.dequeue_task_fair 0.01 ±173% +0.1 0.08 ± 27% perf-profile.children.cycles-pp.write_extent_buffer 0.23 ± 13% +0.1 0.30 ± 10% perf-profile.children.cycles-pp.__kernel_text_address 0.12 ± 33% +0.1 0.19 ± 21% perf-profile.children.cycles-pp.btrfs_leaf_free_space 0.12 ± 5% +0.1 0.20 ± 33% perf-profile.children.cycles-pp.set_extent_buffer_dirty 0.05 ± 60% +0.1 0.13 ± 8% perf-profile.children.cycles-pp.btrfs_free_tree_block 0.25 ± 16% +0.1 0.35 ± 22% perf-profile.children.cycles-pp.btrfs_free_path 0.15 ± 12% +0.1 0.24 ± 25% perf-profile.children.cycles-pp.alloc_extent_buffer 0.20 ± 11% +0.1 0.31 ± 21% perf-profile.children.cycles-pp.memcpy_extent_buffer 0.28 ± 18% +0.1 0.39 ± 15% perf-profile.children.cycles-pp.btrfs_unlock_up_safe 0.10 ± 10% +0.1 0.25 ± 10% perf-profile.children.cycles-pp.btrfs_add_delayed_tree_ref 0.15 ± 28% +0.2 0.31 ± 27% perf-profile.children.cycles-pp.btrfs_add_delayed_data_ref 0.15 ± 30% +0.2 0.31 ± 28% perf-profile.children.cycles-pp.btrfs_inc_extent_ref 0.23 ± 19% +0.2 0.41 ± 25% perf-profile.children.cycles-pp.unpin_extent_cache 0.31 ± 14% +0.2 0.50 ± 23% perf-profile.children.cycles-pp.memmove 0.14 ± 15% +0.2 0.33 ± 15% perf-profile.children.cycles-pp.add_delayed_ref_head 0.94 ± 14% +0.2 1.13 ± 4% perf-profile.children.cycles-pp.schedule 0.26 ± 30% +0.2 0.45 ± 21% perf-profile.children.cycles-pp.__push_leaf_right 0.28 ± 7% +0.2 0.49 ± 22% perf-profile.children.cycles-pp.copy_page 0.27 ± 10% +0.2 0.49 ± 23% perf-profile.children.cycles-pp.copy_extent_buffer_full 0.38 ± 15% +0.2 0.60 ± 17% perf-profile.children.cycles-pp.btrfs_alloc_tree_block 0.54 ± 26% +0.3 0.80 ± 19% perf-profile.children.cycles-pp.btrfs_set_token_32 0.00 +0.5 0.53 ± 56% perf-profile.children.cycles-pp.space_info_add_old_bytes 0.91 ± 15% +0.6 1.50 ± 24% perf-profile.children.cycles-pp.push_leaf_right 1.09 ± 17% +0.6 1.74 ± 25% perf-profile.children.cycles-pp.split_leaf 0.83 ± 10% +0.6 1.48 ± 20% perf-profile.children.cycles-pp.__btrfs_cow_block 0.84 ± 11% +0.7 1.49 ± 19% perf-profile.children.cycles-pp.btrfs_cow_block 1.44 ± 19% +0.9 2.32 ± 26% perf-profile.children.cycles-pp.btrfs_insert_empty_items 1.60 ± 20% +0.9 2.51 ± 22% perf-profile.children.cycles-pp.setup_items_for_insert 2.00 ± 18% +1.2 3.23 ± 27% perf-profile.children.cycles-pp.add_pending_csums 2.00 ± 17% +1.2 3.23 ± 27% perf-profile.children.cycles-pp.btrfs_csum_file_blocks 2.94 ± 23% +1.5 4.45 ± 16% perf-profile.children.cycles-pp.btrfs_duplicate_item 2.37 ± 27% -1.3 1.08 ± 18% perf-profile.self.cycles-pp.__crc32c_le 0.13 ± 46% -0.1 0.03 ±102% perf-profile.self.cycles-pp.memzero_extent_buffer 0.16 ± 25% -0.1 0.08 ± 58% perf-profile.self.cycles-pp.xas_find_marked 0.11 ± 3% +0.0 0.13 ± 3% perf-profile.self.cycles-pp.rb_next 0.06 ± 15% +0.0 0.08 ± 10% perf-profile.self.cycles-pp.dequeue_task_fair 0.07 ± 17% +0.0 0.10 ± 4% perf-profile.self.cycles-pp.rb_erase 0.06 ± 17% +0.0 0.10 ± 30% perf-profile.self.cycles-pp.block_group_cache_tree_search 0.16 ± 6% +0.0 0.19 ± 11% perf-profile.self.cycles-pp.nr_iowait_cpu 0.12 ± 6% +0.0 0.16 ± 11% perf-profile.self.cycles-pp.__switch_to 0.13 ± 14% +0.0 0.17 ± 7% perf-profile.self.cycles-pp.pick_next_task_fair 0.01 ±173% +0.1 0.08 ± 27% perf-profile.self.cycles-pp.write_extent_buffer 0.22 ± 25% +0.1 0.32 ± 13% perf-profile.self.cycles-pp.setup_items_for_insert 0.11 ± 13% +0.2 0.27 ± 15% perf-profile.self.cycles-pp.add_delayed_ref_head 0.31 ± 13% +0.2 0.49 ± 23% perf-profile.self.cycles-pp.memmove 0.27 ± 6% +0.2 0.49 ± 22% perf-profile.self.cycles-pp.copy_page 0.48 ± 24% +0.3 0.74 ± 19% perf-profile.self.cycles-pp.btrfs_set_token_32 fio.write_clat_95__us 600 +-+-------------------------------------------------------------------+ | O O | 550 +-+ O O | | O | 500 +O+ O O O OOOO O O O O O O O OOO | O O O OO O O O O O | 450 +-+ O O O O | | | 400 +-+ | | | 350 +-+ | | + +. .+ + +. .+++ +. .++ .++++.++ +.++| 300 +-+++ +.++++.+++ ++++ + +.+++ ++++ +.+++ ++++ ++ + | | | 250 +-+-------------------------------------------------------------------+ fio.write_clat_99__us 1400 +-+------------------------------------------------------------------+ 1300 +-+ O O | | O O O OO O O | 1200 +-+ O O O O O O O | 1100 OOO O O O O O O O OO O O O | | O O O | 1000 +-+ O O O | 900 +-+ | 800 +-+ | | | 700 +-+ | 600 +-+ | | .+ | 500 +-+. + .+ + .+ .++ +. + .+ +.+ + + +. +++.++++ +++.++| 400 +-+------------------------------------------------------------------+ fio.latency_50us_ 40 +-+--------------------------------------------------------------------+ | | 35 +O+ OO O O O OO O O O | 30 O-+OOO O OO OO O O O O O O OO O | | O O O O O O | 25 +-+ O O | | | 20 +-+ | | | 15 +-+ | 10 +-+ | | | 5 +-+ | | | 0 +-+--------------------------------------------------------------------+ fio.latency_100us_ 30 +-+--------------------------------------------------------------------+ | O O | 25 +-+ O O O O O O O OO | O O O O O OO O OO | |O OOO OO OO O OOO OO O | 20 +-+ O | | | 15 +-+ | | | 10 +-+ | | | | | 5 +-+ | |+.++++.++++.++++.+++.++++.++++.+++.++++.++++.++++.+++.++++.++++.++++.+| 0 +-+--------------------------------------------------------------------+ fio.latency_250us_ 80 +-+--------------------------------------------------------------------+ |+.+ ++.++ +.++ +.+++ +++.+++ +++.+ ++.++++ +++.+ +.+ ++.++++.++++.+| 70 +-+ | | | | | 60 +-+ | | | 50 +-+ | | | 40 +-+ | | | | | 30 +-+ O O O O O O O | OO OO O OOO O OOO O O O O OO O OOO O | 20 +-+------------OO-----------------O--OOO-O-----------------------------+ fio.latency_750us_ 4 +-+-------------------------------------------------------------------+ | O O | 3.5 +-+ | 3 +-+ O O O | |O O OOOO O O O OO O O O O | 2.5 O-+O OO OO O O O | | OO OO O OO O | 2 +-+ | | | 1.5 +-+ | 1 +-+ | | | 0.5 +-+ | |+.++++.++ +.++++.++++.++++.++++.++++.++++.++++.++++.++++.++++.++++.++| 0 +-+-------------------------------------------------------------------+ fio.latency_1000us_ 1.4 +-+-------------------------------------------------------------------+ | O O | 1.2 +-+ | | O O | 1 +O+ O O O OO O O O O | O O O O OO O O O O O O O | 0.8 +-+ O O O O O O | | O O O O | 0.6 +-+ | | | 0.4 +-+ | | | 0.2 +-++ +. + .+ + ++. + .+ + ++. + .+ + ++ .++++.++++.++| |+ +++.+++ + ++ + +.++ + ++ + +.++ + ++ + +.+ + | 0 +-+-------------------------------------------------------------------+ fio.latency_4ms_ 0.35 +-+------------------------------------------------------------------+ | | 0.3 +-+ O O O | | OO O OOO O O O O O OO | 0.25 +-O O OO O OO | OO O O O O O | 0.2 +-+ O O O O | | O O O | 0.15 +-+ | | | 0.1 +-+ | | | 0.05 +-+ | |++.++++.++++.+++++.++++.++++.++++.+++++.++++.++++.+++++.++++.++++.++| 0 +-+------------------------------------------------------------------+ fio.latency_10ms_ 0.14 +-+------------------------------------------------------------------+ | | 0.12 +-+ O O O O | | O O OO O O | 0.1 +-+ O O | O O O O O O O OOOO O O O | 0.08 +O+ O O O O O O O | | O O O | 0.06 +-+ | | | 0.04 +-+ | | + | 0.02 +-+ +.+ ++. +++.++ +.++++.++| |++.++++.++++.+++++.+++ +++.++++.+++ ++++.++++.++ | 0 +-+------------------------------------------------------------------+ fio.latency_20ms_ 0.03 +-+------O----------------------------------------------------------+ 0.028 +-+ | | O O O O | 0.026 +-+ O | 0.024 +-O O O | | O O O | 0.022 O-+ O O O | 0.02 +-+ O O O OO OO | 0.018 +-+ O O O O OO O | |O O | 0.016 +-+ O O O O O | 0.014 +-+ | | | 0.012 +-+ + | 0.01 +-+-----------------------------------------------------------------+ [*] bisect-good sample [O] bisect-bad sample 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, Rong Chen