| Message ID | 20220425020926.1264611-1-longman@redhat.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | cgroup/cpuset: Remove redundant cpu/node masks setup in cpuset_init_smp() | expand |
Hi Waiman, Thanks for the patch! On Sun, Apr 24, 2022 at 10:09:26PM -0400, Waiman Long wrote: > There are 3 places where the cpu and node masks of the top cpuset can > be initialized in the order they are executed: > 1) start_kernel -> cpuset_init() > 2) start_kernel -> cgroup_init() -> cpuset_bind() > 3) kernel_init_freeable() -> do_basic_setup() -> cpuset_init_smp() > > The first cpuset_init() function just sets all the bits in the masks. > The last one executed is cpuset_init_smp() which sets up cpu and node > masks suitable for v1, but not v2. cpuset_bind() does the right setup > for both v1 and v2 assuming that effective_mems and effective_cpus have > been set up properly which is not strictly the case here. As a result, > cpu and memory node hot add may fail to update the cpu and node masks > of the top cpuset to include the newly added cpu or node in a cgroup > v2 environment. > > To fix this problem, the redundant cpus_allowed and mems_allowed > mask setup in cpuset_init_smp() are removed. The effective_cpus and > effective_mems setup there are moved to cpuset_bind(). > > cc: stable@vger.kernel.org > Signed-off-by: Waiman Long <longman@redhat.com> > --- > kernel/cgroup/cpuset.c | 10 +++------- > 1 file changed, 3 insertions(+), 7 deletions(-) > > diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c > index 9390bfd9f1cd..a2e15a43397e 100644 > --- a/kernel/cgroup/cpuset.c > +++ b/kernel/cgroup/cpuset.c > @@ -2961,6 +2961,9 @@ static void cpuset_bind(struct cgroup_subsys_state *root_css) > percpu_down_write(&cpuset_rwsem); > spin_lock_irq(&callback_lock); > > + cpumask_copy(top_cpuset.effective_cpus, cpu_active_mask); > + top_cpuset.effective_mems = node_states[N_MEMORY]; > + > if (is_in_v2_mode()) { > cpumask_copy(top_cpuset.cpus_allowed, cpu_possible_mask); > top_cpuset.mems_allowed = node_possible_map; > @@ -3390,13 +3393,6 @@ static struct notifier_block cpuset_track_online_nodes_nb = { > */ > void __init cpuset_init_smp(void) > { > - cpumask_copy(top_cpuset.cpus_allowed, cpu_active_mask); > - top_cpuset.mems_allowed = node_states[N_MEMORY]; > - top_cpuset.old_mems_allowed = top_cpuset.mems_allowed; > - > - cpumask_copy(top_cpuset.effective_cpus, cpu_active_mask); > - top_cpuset.effective_mems = node_states[N_MEMORY]; IIUC, the init order is: cpuset_bind() smp_init() cpuset_init_smp() while all cpus except boot cpu is brought up in smp_init(), so I'm thinking moving the cpus_allowed init from cpuset_init_smp() to cpuset_bind() may cause some problem. Thanks, Feng
On 4/25/22 03:30, Feng Tang wrote: > Hi Waiman, > > Thanks for the patch! > > On Sun, Apr 24, 2022 at 10:09:26PM -0400, Waiman Long wrote: >> There are 3 places where the cpu and node masks of the top cpuset can >> be initialized in the order they are executed: >> 1) start_kernel -> cpuset_init() >> 2) start_kernel -> cgroup_init() -> cpuset_bind() >> 3) kernel_init_freeable() -> do_basic_setup() -> cpuset_init_smp() >> >> The first cpuset_init() function just sets all the bits in the masks. >> The last one executed is cpuset_init_smp() which sets up cpu and node >> masks suitable for v1, but not v2. cpuset_bind() does the right setup >> for both v1 and v2 assuming that effective_mems and effective_cpus have >> been set up properly which is not strictly the case here. As a result, >> cpu and memory node hot add may fail to update the cpu and node masks >> of the top cpuset to include the newly added cpu or node in a cgroup >> v2 environment. >> >> To fix this problem, the redundant cpus_allowed and mems_allowed >> mask setup in cpuset_init_smp() are removed. The effective_cpus and >> effective_mems setup there are moved to cpuset_bind(). >> >> cc: stable@vger.kernel.org >> Signed-off-by: Waiman Long <longman@redhat.com> >> --- >> kernel/cgroup/cpuset.c | 10 +++------- >> 1 file changed, 3 insertions(+), 7 deletions(-) >> >> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c >> index 9390bfd9f1cd..a2e15a43397e 100644 >> --- a/kernel/cgroup/cpuset.c >> +++ b/kernel/cgroup/cpuset.c >> @@ -2961,6 +2961,9 @@ static void cpuset_bind(struct cgroup_subsys_state *root_css) >> percpu_down_write(&cpuset_rwsem); >> spin_lock_irq(&callback_lock); >> >> + cpumask_copy(top_cpuset.effective_cpus, cpu_active_mask); >> + top_cpuset.effective_mems = node_states[N_MEMORY]; >> + >> if (is_in_v2_mode()) { >> cpumask_copy(top_cpuset.cpus_allowed, cpu_possible_mask); >> top_cpuset.mems_allowed = node_possible_map; >> @@ -3390,13 +3393,6 @@ static struct notifier_block cpuset_track_online_nodes_nb = { >> */ >> void __init cpuset_init_smp(void) >> { >> - cpumask_copy(top_cpuset.cpus_allowed, cpu_active_mask); >> - top_cpuset.mems_allowed = node_states[N_MEMORY]; >> - top_cpuset.old_mems_allowed = top_cpuset.mems_allowed; >> - >> - cpumask_copy(top_cpuset.effective_cpus, cpu_active_mask); >> - top_cpuset.effective_mems = node_states[N_MEMORY]; > IIUC, the init order is: > cpuset_bind() > smp_init() > cpuset_init_smp() > > while all cpus except boot cpu is brought up in smp_init(), so I'm > thinking moving the cpus_allowed init from cpuset_init_smp() to > cpuset_bind() may cause some problem. Good point. So cpuset_init_smp() is still useful for setting the right effective_cpus and probably effective_mems. I will update the patch accordingly. Thanks, Longman > > Thanks, > Feng >
Greeting,
FYI, we noticed a 2583.0% regression of kbuild.buildtime_per_iteration due to commit:
commit: 7aaa45f8374887d98cd479a6bdd3b4c74e6a981a ("[PATCH] cgroup/cpuset: Remove redundant cpu/node masks setup in cpuset_init_smp()")
url: https://github.com/intel-lab-lkp/linux/commits/Waiman-Long/cgroup-cpuset-Remove-redundant-cpu-node-masks-setup-in-cpuset_init_smp/20220425-101126
base: https://git.kernel.org/cgit/linux/kernel/git/tj/cgroup.git for-next
patch link: https://lore.kernel.org/linux-mm/20220425020926.1264611-1-longman@redhat.com
in testcase: kbuild
on test machine: 88 threads 2 sockets Intel(R) Xeon(R) Gold 6238M CPU @ 2.10GHz with 128G memory
with following parameters:
memory.high: 90%
memory.low: 50%
memory.max: max
pids.max: 10000
runtime: 300s
nr_task: 50%
target: autoksyms_recursive
cpufreq_governor: performance
ucode: 0x500320a
If you fix the issue, kindly add following tag
Reported-by: kernel test robot <oliver.sang@intel.com>
Details are as below:
-------------------------------------------------------------------------------------------------->
To reproduce:
git clone https://github.com/intel/lkp-tests.git
cd lkp-tests
sudo bin/lkp install job.yaml # job file is attached in this email
bin/lkp split-job --compatible job.yaml # generate the yaml file for lkp run
sudo bin/lkp run generated-yaml-file
# if come across any failure that blocks the test,
# please remove ~/.lkp and /lkp dir to run from a clean state.
=========================================================================================
compiler/cpufreq_governor/kconfig/memory.high/memory.low/memory.max/nr_task/pids.max/rootfs/runtime/target/tbox_group/testcase/ucode:
gcc-11/performance/x86_64-rhel-8.3/90%/50%/max/50%/10000/debian-10.4-x86_64-20200603.cgz/300s/autoksyms_recursive/lkp-csl-2sp9/kbuild/0x500320a
commit:
4ab93063c8 ("cgroup: Add test_cpucg_weight_underprovisioned() testcase")
7aaa45f837 ("cgroup/cpuset: Remove redundant cpu/node masks setup in cpuset_init_smp()")
4ab93063c83a2478 7aaa45f8374887d98cd479a6bdd
---------------- ---------------------------
%stddev %change %stddev
\ | \
44.43 +2583.0% 1192 kbuild.buildtime_per_iteration
311.14 +283.0% 1191 kbuild.time.elapsed_time
311.14 +283.0% 1191 kbuild.time.elapsed_time.max
363486 -4.7% 346532 kbuild.time.involuntary_context_switches
7528 ± 2% -99.1% 68.00 kbuild.time.major_page_faults
3.854e+08 -85.7% 55047165 kbuild.time.minor_page_faults
2608 -96.4% 95.00 kbuild.time.percent_of_cpu_this_job_got
1361 -89.5% 142.46 kbuild.time.system_time
6756 -85.3% 991.32 kbuild.time.user_time
1763962 -84.6% 270814 kbuild.time.voluntary_context_switches
7.00 -85.7% 1.00 kbuild.workload
1.886e+10 +439.9% 1.018e+11 cpuidle..time
40524113 +397.6% 2.016e+08 ± 9% cpuidle..usage
352.84 +249.7% 1233 uptime.boot
22088 +378.4% 105677 uptime.idle
68.86 +28.8 97.63 mpstat.cpu.all.idle%
0.16 ± 3% -0.1 0.08 ± 8% mpstat.cpu.all.soft%
5.23 -5.1 0.15 mpstat.cpu.all.sys%
24.48 -23.5 0.96 mpstat.cpu.all.usr%
68.83 +42.1% 97.83 vmstat.cpu.id
24.00 -100.0% 0.00 vmstat.cpu.us
26.00 +73.1% 45.00 vmstat.procs.r
17438 -88.3% 2045 vmstat.system.cs
1.749e+08 -70.6% 51378800 numa-numastat.node0.local_node
1.75e+08 -70.6% 51378813 numa-numastat.node0.numa_hit
50974 ± 50% -100.0% 13.50 ± 75% numa-numastat.node0.other_node
1.75e+08 -100.0% 82.83 ±223% numa-numastat.node1.local_node
1.75e+08 -100.0% 79394 numa-numastat.node1.numa_hit
28316 ± 91% +180.1% 79311 numa-numastat.node1.other_node
6700 ± 4% +69.9% 11385 ± 3% meminfo.Active
6540 ± 4% +69.1% 11063 meminfo.Active(anon)
223421 +31.7% 294262 ± 2% meminfo.AnonHugePages
1300267 +31.4% 1708081 meminfo.AnonPages
1576268 +25.7% 1981479 meminfo.Committed_AS
1338713 +28.5% 1719792 meminfo.Inactive
1338565 +28.4% 1719004 meminfo.Inactive(anon)
18332 -11.0% 16310 meminfo.KernelStack
93129 -43.8% 52361 meminfo.Mapped
14656 ± 2% -8.9% 13355 meminfo.PageTables
43576 ± 3% -52.0% 20896 meminfo.Shmem
7581989 -10.5% 6785067 meminfo.max_used_kB
875.33 -100.0% 0.00 turbostat.Avg_MHz
31.74 -31.7 0.00 turbostat.Busy%
2764 -100.0% 0.00 turbostat.Bzy_MHz
173892 ± 10% -100.0% 0.00 turbostat.C1
0.02 ± 17% -0.0 0.00 turbostat.C1%
33508125 ± 20% -100.0% 0.00 turbostat.C1E
48.18 ± 44% -48.2 0.00 turbostat.C1E%
6776884 ±102% -100.0% 0.00 turbostat.C6
20.30 ±104% -20.3 0.00 turbostat.C6%
66.40 ± 4% -100.0% 0.00 turbostat.CPU%c1
1.86 ±139% -100.0% 0.00 turbostat.CPU%c6
59.17 -100.0% 0.00 turbostat.CoreTmp
0.33 -100.0% 0.00 turbostat.IPC
55232878 -100.0% 0.00 turbostat.IRQ
56380 ± 6% -100.0% 0.00 turbostat.POLL
59.00 -100.0% 0.00 turbostat.PkgTmp
215.76 -100.0% 0.00 turbostat.PkgWatt
16.16 -100.0% 0.00 turbostat.RAMWatt
2095 -100.0% 0.00 turbostat.TSC_MHz
813.00 ± 11% +237.1% 2740 numa-vmstat.node0.nr_active_anon
183776 +131.6% 425651 numa-vmstat.node0.nr_anon_pages
186417 +129.7% 428211 numa-vmstat.node0.nr_inactive_anon
1978 +68.0% 3324 numa-vmstat.node0.nr_page_table_pages
3282 ± 7% +53.1% 5023 numa-vmstat.node0.nr_shmem
813.00 ± 11% +237.1% 2740 numa-vmstat.node0.nr_zone_active_anon
186416 +129.7% 428211 numa-vmstat.node0.nr_zone_inactive_anon
1.75e+08 -70.6% 51378302 numa-vmstat.node0.numa_hit
1.749e+08 -70.6% 51378288 numa-vmstat.node0.numa_local
50973 ± 50% -100.0% 13.50 ± 75% numa-vmstat.node0.numa_other
823.33 ± 5% -97.5% 20.33 ± 12% numa-vmstat.node1.nr_active_anon
139750 ± 2% -99.9% 88.17 ±114% numa-vmstat.node1.nr_anon_pages
157716 ±183% -97.6% 3733 ±208% numa-vmstat.node1.nr_file_pages
146649 ± 2% -99.8% 265.00 ± 56% numa-vmstat.node1.nr_inactive_anon
8288 ± 4% -31.8% 5653 ± 5% numa-vmstat.node1.nr_kernel_stack
8789 ± 47% -99.6% 32.00 ±120% numa-vmstat.node1.nr_mapped
1658 -99.8% 4.00 ± 73% numa-vmstat.node1.nr_page_table_pages
7610 ± 3% -97.4% 197.17 ± 27% numa-vmstat.node1.nr_shmem
11225 ±101% -74.3% 2883 ± 9% numa-vmstat.node1.nr_slab_reclaimable
20034 ± 3% -25.8% 14861 ± 4% numa-vmstat.node1.nr_slab_unreclaimable
823.33 ± 5% -97.5% 20.33 ± 12% numa-vmstat.node1.nr_zone_active_anon
146648 ± 2% -99.8% 265.00 ± 56% numa-vmstat.node1.nr_zone_inactive_anon
1.75e+08 -100.0% 79394 numa-vmstat.node1.numa_hit
1.75e+08 -100.0% 82.83 ±223% numa-vmstat.node1.numa_local
28312 ± 91% +180.1% 79311 numa-vmstat.node1.numa_other
1642 ± 4% +68.3% 2763 proc-vmstat.nr_active_anon
322313 +32.5% 427139 proc-vmstat.nr_anon_pages
864421 -1.5% 851593 proc-vmstat.nr_file_pages
332108 +29.4% 429879 proc-vmstat.nr_inactive_anon
18183 -10.3% 16310 proc-vmstat.nr_kernel_stack
23431 -43.6% 13208 proc-vmstat.nr_mapped
3550 ± 2% -6.0% 3337 proc-vmstat.nr_page_table_pages
11164 ± 3% -53.2% 5223 proc-vmstat.nr_shmem
46151 -2.2% 45116 proc-vmstat.nr_slab_reclaimable
46770 -8.6% 42749 proc-vmstat.nr_slab_unreclaimable
1642 ± 4% +68.3% 2763 proc-vmstat.nr_zone_active_anon
332108 +29.4% 429879 proc-vmstat.nr_zone_inactive_anon
360277 ± 30% -99.8% 680.17 ±112% proc-vmstat.numa_hint_faults
357678 ± 31% -99.8% 680.17 ±112% proc-vmstat.numa_hint_faults_local
3.5e+08 -85.3% 51462030 proc-vmstat.numa_hit
3.499e+08 -85.3% 51382705 proc-vmstat.numa_local
10372 ±129% -100.0% 0.00 proc-vmstat.numa_pages_migrated
790289 ± 15% -75.5% 193877 ± 7% proc-vmstat.numa_pte_updates
290045 -86.1% 40379 proc-vmstat.pgactivate
3.498e+08 -85.3% 51462440 proc-vmstat.pgalloc_normal
3.871e+08 -85.1% 57578373 proc-vmstat.pgfault
3.497e+08 -85.3% 51397471 proc-vmstat.pgfree
10372 ±129% -100.0% 0.00 proc-vmstat.pgmigrate_success
2920438 -78.6% 624274 proc-vmstat.pgreuse
18128 -85.7% 2594 proc-vmstat.thp_fault_alloc
771909 -85.7% 110390 proc-vmstat.unevictable_pgs_culled
3411 ± 11% +226.2% 11127 numa-meminfo.node0.Active
3251 ± 11% +237.3% 10966 numa-meminfo.node0.Active(anon)
188175 ± 6% +56.0% 293523 ± 2% numa-meminfo.node0.AnonHugePages
737315 +130.8% 1701834 numa-meminfo.node0.AnonPages
1657848 ± 5% +85.9% 3081876 ± 4% numa-meminfo.node0.AnonPages.max
748023 +129.0% 1712859 numa-meminfo.node0.Inactive
747875 +128.9% 1712108 numa-meminfo.node0.Inactive(anon)
4306840 ± 28% +35.6% 5839514 numa-meminfo.node0.MemUsed
7772 ± 3% +71.1% 13297 numa-meminfo.node0.PageTables
13182 ± 8% +52.5% 20099 numa-meminfo.node0.Shmem
3307 ± 5% -92.7% 242.00 ±147% numa-meminfo.node1.Active
3307 ± 5% -97.5% 81.33 ± 12% numa-meminfo.node1.Active(anon)
34918 ± 34% -100.0% 0.00 numa-meminfo.node1.AnonHugePages
557333 ± 2% -99.9% 353.33 ±114% numa-meminfo.node1.AnonPages
1876230 ± 3% -100.0% 496.67 ± 96% numa-meminfo.node1.AnonPages.max
631836 ±183% -97.6% 14935 ±208% numa-meminfo.node1.FilePages
586137 ± 2% -99.8% 1096 ± 51% numa-meminfo.node1.Inactive
586137 ± 2% -99.8% 1061 ± 56% numa-meminfo.node1.Inactive(anon)
44909 ±101% -74.3% 11533 ± 9% numa-meminfo.node1.KReclaimable
8244 ± 4% -31.4% 5653 ± 5% numa-meminfo.node1.KernelStack
35330 ± 46% -99.6% 129.17 ±119% numa-meminfo.node1.Mapped
1612958 ± 74% -79.4% 333010 ± 7% numa-meminfo.node1.MemUsed
6577 -99.8% 16.17 ± 73% numa-meminfo.node1.PageTables
44909 ±101% -74.3% 11533 ± 9% numa-meminfo.node1.SReclaimable
80135 ± 3% -25.8% 59446 ± 4% numa-meminfo.node1.SUnreclaim
31315 ± 3% -97.5% 789.17 ± 27% numa-meminfo.node1.Shmem
125044 ± 38% -43.2% 70980 ± 4% numa-meminfo.node1.Slab
1.978e+10 -95.6% 8.797e+08 perf-stat.i.branch-instructions
5.616e+08 -95.6% 24754884 ± 8% perf-stat.i.branch-misses
1.495e+08 -94.7% 7868663 ± 5% perf-stat.i.cache-misses
6.176e+08 -94.5% 34193232 ± 36% perf-stat.i.cache-references
17499 -88.4% 2029 perf-stat.i.context-switches
0.93 ± 3% +36.6% 1.27 ± 5% perf-stat.i.cpi
7.611e+10 -93.3% 5.121e+09 ± 4% perf-stat.i.cpu-cycles
1275 -100.0% 0.04 ± 28% perf-stat.i.cpu-migrations
1029 ± 5% -31.3% 707.28 ± 2% perf-stat.i.cycles-between-cache-misses
14172722 -94.1% 836284 ± 47% perf-stat.i.dTLB-load-misses
2.503e+10 -95.5% 1.122e+09 perf-stat.i.dTLB-loads
7917427 -95.8% 332811 ± 15% perf-stat.i.dTLB-store-misses
1.244e+10 -95.4% 5.666e+08 perf-stat.i.dTLB-stores
24631063 -92.3% 1894591 ± 2% perf-stat.i.iTLB-load-misses
71323368 -94.9% 3665231 ± 6% perf-stat.i.iTLB-loads
9.005e+10 -95.5% 4.054e+09 perf-stat.i.instructions
3571 -39.7% 2152 perf-stat.i.instructions-per-iTLB-miss
1.10 ± 2% -28.1% 0.79 ± 5% perf-stat.i.ipc
24.38 ± 2% -99.5% 0.11 ± 23% perf-stat.i.major-faults
0.86 -93.3% 0.06 ± 4% perf-stat.i.metric.GHz
658.74 -95.6% 29.19 perf-stat.i.metric.M/sec
1236978 -96.1% 48019 perf-stat.i.minor-faults
24.93 ± 5% -18.7 6.26 ± 22% perf-stat.i.node-load-miss-rate%
6765575 -97.4% 178075 ± 26% perf-stat.i.node-load-misses
55797319 -94.4% 3120263 perf-stat.i.node-loads
18.32 ± 4% -9.7 8.62 ± 22% perf-stat.i.node-store-miss-rate%
1648658 -97.0% 50137 ± 26% perf-stat.i.node-store-misses
11762943 -95.0% 585083 perf-stat.i.node-stores
1237003 -96.1% 48019 perf-stat.i.page-faults
6.86 +23.4% 8.46 ± 38% perf-stat.overall.MPKI
0.85 +49.5% 1.26 ± 5% perf-stat.overall.cpi
508.96 +27.9% 651.10 perf-stat.overall.cycles-between-cache-misses
25.67 +8.5 34.14 ± 3% perf-stat.overall.iTLB-load-miss-rate%
3655 -41.4% 2140 perf-stat.overall.instructions-per-iTLB-miss
1.18 -32.9% 0.79 ± 5% perf-stat.overall.ipc
10.82 -5.4 5.38 ± 24% perf-stat.overall.node-load-miss-rate%
12.30 -4.4 7.85 ± 23% perf-stat.overall.node-store-miss-rate%
4.002e+12 +20.6% 4.826e+12 perf-stat.overall.path-length
1.972e+10 -95.5% 8.791e+08 perf-stat.ps.branch-instructions
5.599e+08 -95.6% 24737459 ± 8% perf-stat.ps.branch-misses
1.491e+08 -94.7% 7862895 ± 5% perf-stat.ps.cache-misses
6.158e+08 -94.5% 34168138 ± 36% perf-stat.ps.cache-references
17447 -88.4% 2027 perf-stat.ps.context-switches
7.588e+10 -93.3% 5.118e+09 ± 4% perf-stat.ps.cpu-cycles
1272 -100.0% 0.04 ± 28% perf-stat.ps.cpu-migrations
14129808 -94.1% 835661 ± 47% perf-stat.ps.dTLB-load-misses
2.495e+10 -95.5% 1.121e+09 perf-stat.ps.dTLB-loads
7893270 -95.8% 332570 ± 15% perf-stat.ps.dTLB-store-misses
1.24e+10 -95.4% 5.662e+08 perf-stat.ps.dTLB-stores
24556703 -92.3% 1893249 ± 2% perf-stat.ps.iTLB-load-misses
71106326 -94.8% 3662587 ± 6% perf-stat.ps.iTLB-loads
8.978e+10 -95.5% 4.051e+09 perf-stat.ps.instructions
24.32 ± 2% -99.5% 0.11 ± 24% perf-stat.ps.major-faults
1233214 -96.1% 47986 perf-stat.ps.minor-faults
6745825 -97.4% 177950 ± 26% perf-stat.ps.node-load-misses
55627061 -94.4% 3117959 perf-stat.ps.node-loads
1644068 -97.0% 50102 ± 26% perf-stat.ps.node-store-misses
11727184 -95.0% 584651 perf-stat.ps.node-stores
1233238 -96.1% 47986 perf-stat.ps.page-faults
2.802e+13 -82.8% 4.826e+12 perf-stat.total.instructions
0.39 +3586.3% 14.32 ± 4% sched_debug.cfs_rq:/.h_nr_running.avg
1.11 ± 7% +4138.2% 47.09 sched_debug.cfs_rq:/.h_nr_running.max
0.43 +4683.1% 20.72 sched_debug.cfs_rq:/.h_nr_running.stddev
14081 ± 20% +2107.0% 310782 ± 4% sched_debug.cfs_rq:/.load.avg
407170 ± 49% +170.4% 1101004 sched_debug.cfs_rq:/.load.max
52684 ± 42% +768.5% 457567 sched_debug.cfs_rq:/.load.stddev
37.01 ± 46% +723.1% 304.63 ± 4% sched_debug.cfs_rq:/.load_avg.avg
144.31 ± 45% +210.7% 448.32 sched_debug.cfs_rq:/.load_avg.stddev
2488735 -92.5% 187680 ± 3% sched_debug.cfs_rq:/.min_vruntime.avg
2938975 -80.2% 581666 sched_debug.cfs_rq:/.min_vruntime.max
2060745 ± 2% -99.5% 10250 ± 34% sched_debug.cfs_rq:/.min_vruntime.min
182406 ± 6% +39.2% 253824 sched_debug.cfs_rq:/.min_vruntime.stddev
0.39 -23.7% 0.30 ± 4% sched_debug.cfs_rq:/.nr_running.avg
400.60 +3272.5% 13510 ± 5% sched_debug.cfs_rq:/.runnable_avg.avg
971.78 ± 2% +4430.1% 44022 sched_debug.cfs_rq:/.runnable_avg.max
370.10 +5173.5% 19517 sched_debug.cfs_rq:/.runnable_avg.stddev
12579 ±1058% -3228.1% -393486 sched_debug.cfs_rq:/.spread0.avg
462824 ± 24% -99.9% 540.01 ± 92% sched_debug.cfs_rq:/.spread0.max
182401 ± 6% +39.2% 253852 sched_debug.cfs_rq:/.spread0.stddev
400.13 -25.0% 300.07 ± 4% sched_debug.cfs_rq:/.util_avg.avg
369.97 +18.4% 437.88 sched_debug.cfs_rq:/.util_avg.stddev
10.90 ± 10% -87.5% 1.37 ± 48% sched_debug.cfs_rq:/.util_est_enqueued.avg
454.44 ± 14% -90.3% 44.23 ± 3% sched_debug.cfs_rq:/.util_est_enqueued.max
60.82 ± 13% -90.9% 5.53 ± 14% sched_debug.cfs_rq:/.util_est_enqueued.stddev
0.92 ± 36% -90.0% 0.09 ±176% sched_debug.cfs_rq:/init.scope.load_avg.avg
1.08 ± 17% -91.5% 0.09 ±176% sched_debug.cfs_rq:/init.scope.load_avg.max
0.50 ±107% +2506.1% 12.97 ± 57% sched_debug.cfs_rq:/init.scope.min_vruntime.avg
0.54 ±104% +2288.9% 13.00 ± 57% sched_debug.cfs_rq:/init.scope.min_vruntime.max
0.45 ±111% +2768.6% 12.93 ± 57% sched_debug.cfs_rq:/init.scope.min_vruntime.min
0.92 ± 36% -90.0% 0.09 ±176% sched_debug.cfs_rq:/init.scope.runnable_avg.avg
1.08 ± 17% -91.5% 0.09 ±176% sched_debug.cfs_rq:/init.scope.runnable_avg.max
0.51 ±123% +1809.6% 9.78 ± 40% sched_debug.cfs_rq:/init.scope.se->sum_exec_runtime.avg
0.52 ±122% +1769.0% 9.78 ± 40% sched_debug.cfs_rq:/init.scope.se->sum_exec_runtime.max
0.50 ±126% +1852.0% 9.77 ± 40% sched_debug.cfs_rq:/init.scope.se->sum_exec_runtime.min
927421 ±116% -88.6% 105436 ± 29% sched_debug.cfs_rq:/init.scope.se->vruntime.avg
927672 ±116% -88.6% 105945 ± 29% sched_debug.cfs_rq:/init.scope.se->vruntime.max
927171 ±117% -88.7% 104926 ± 29% sched_debug.cfs_rq:/init.scope.se->vruntime.min
-903221 -88.3% -106040 sched_debug.cfs_rq:/init.scope.spread0.avg
-903221 -88.3% -106040 sched_debug.cfs_rq:/init.scope.spread0.max
-903221 -88.3% -106040 sched_debug.cfs_rq:/init.scope.spread0.min
1.08 ± 17% -91.5% 0.09 ±176% sched_debug.cfs_rq:/init.scope.tg_load_avg.avg
1.08 ± 17% -91.5% 0.09 ±176% sched_debug.cfs_rq:/init.scope.tg_load_avg.max
1.08 ± 17% -91.5% 0.09 ±176% sched_debug.cfs_rq:/init.scope.tg_load_avg.min
0.92 ± 36% -90.0% 0.09 ±176% sched_debug.cfs_rq:/init.scope.tg_load_avg_contrib.avg
1.08 ± 17% -91.5% 0.09 ±176% sched_debug.cfs_rq:/init.scope.tg_load_avg_contrib.max
0.92 ± 36% -96.4% 0.03 ±111% sched_debug.cfs_rq:/init.scope.util_avg.avg
1.08 ± 17% -96.9% 0.03 ±111% sched_debug.cfs_rq:/init.scope.util_avg.max
109.53 ±107% +50485.6% 55405 sched_debug.cfs_rq:/kbuild.1.MIN_vruntime.avg
9638 ±107% +475.0% 55414 sched_debug.cfs_rq:/kbuild.1.MIN_vruntime.max
-0.30 -1.9e+07% 55396 sched_debug.cfs_rq:/kbuild.1.MIN_vruntime.min
0.39 ± 2% +11812.2% 46.70 sched_debug.cfs_rq:/kbuild.1.h_nr_running.avg
1.00 +4594.2% 46.94 sched_debug.cfs_rq:/kbuild.1.h_nr_running.max
410733 ± 2% +11823.3% 48972868 sched_debug.cfs_rq:/kbuild.1.load.avg
1048576 +4594.2% 49221905 sched_debug.cfs_rq:/kbuild.1.load.max
659.13 ± 3% +7394.5% 49398 sched_debug.cfs_rq:/kbuild.1.load_avg.avg
2647 ± 11% +1777.5% 49713 sched_debug.cfs_rq:/kbuild.1.load_avg.max
0.03 ±223% +1.8e+08% 49083 ± 2% sched_debug.cfs_rq:/kbuild.1.load_avg.min
109.53 ±107% +50489.2% 55409 sched_debug.cfs_rq:/kbuild.1.max_vruntime.avg
9638 ±107% +475.0% 55418 sched_debug.cfs_rq:/kbuild.1.max_vruntime.max
-0.30 -1.9e+07% 55400 sched_debug.cfs_rq:/kbuild.1.max_vruntime.min
87264 -36.5% 55405 sched_debug.cfs_rq:/kbuild.1.min_vruntime.avg
99270 -44.2% 55414 sched_debug.cfs_rq:/kbuild.1.min_vruntime.max
75767 -26.9% 55396 sched_debug.cfs_rq:/kbuild.1.min_vruntime.min
5024 ± 7% -99.8% 9.08 ±100% sched_debug.cfs_rq:/kbuild.1.min_vruntime.stddev
0.39 ± 2% +11812.2% 46.70 sched_debug.cfs_rq:/kbuild.1.nr_running.avg
1.00 +4594.2% 46.94 sched_debug.cfs_rq:/kbuild.1.nr_running.max
35.29 ± 22% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.load_avg.avg
919.11 ± 14% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.load_avg.max
156.52 ± 17% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.load_avg.stddev
14.27 ± 27% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.runnable_avg.avg
419.11 ± 25% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.runnable_avg.max
65.99 ± 27% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.runnable_avg.stddev
14.27 ± 27% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.util_avg.avg
419.11 ± 25% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.util_avg.max
65.99 ± 27% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.util_avg.stddev
412.27 +10527.3% 43813 sched_debug.cfs_rq:/kbuild.1.runnable_avg.avg
967.14 ± 2% +4443.7% 43944 sched_debug.cfs_rq:/kbuild.1.runnable_avg.max
0.03 ±223% +1.6e+08% 43681 sched_debug.cfs_rq:/kbuild.1.runnable_avg.min
368.39 -64.4% 131.08 ±101% sched_debug.cfs_rq:/kbuild.1.runnable_avg.stddev
13.05 ± 10% +7646.3% 1011 sched_debug.cfs_rq:/kbuild.1.se->avg.load_avg.avg
87.56 ± 16% +1069.5% 1023 sched_debug.cfs_rq:/kbuild.1.se->avg.load_avg.max
412.13 +10524.7% 43787 sched_debug.cfs_rq:/kbuild.1.se->avg.runnable_avg.avg
967.47 ± 2% +4439.4% 43917 sched_debug.cfs_rq:/kbuild.1.se->avg.runnable_avg.max
0.03 ±223% +1.6e+08% 43657 sched_debug.cfs_rq:/kbuild.1.se->avg.runnable_avg.min
368.51 -64.7% 129.95 ±101% sched_debug.cfs_rq:/kbuild.1.se->avg.runnable_avg.stddev
411.92 +141.6% 995.10 sched_debug.cfs_rq:/kbuild.1.se->avg.util_avg.avg
0.03 ±223% +3.5e+06% 983.40 ± 2% sched_debug.cfs_rq:/kbuild.1.se->avg.util_avg.min
368.41 -96.8% 11.70 ±100% sched_debug.cfs_rq:/kbuild.1.se->avg.util_avg.stddev
188343 +220.5% 603616 sched_debug.cfs_rq:/kbuild.1.se->exec_start.avg
188644 +220.2% 604099 sched_debug.cfs_rq:/kbuild.1.se->exec_start.max
186221 +223.9% 603133 sched_debug.cfs_rq:/kbuild.1.se->exec_start.min
51889 ± 7% +1920.8% 1048576 sched_debug.cfs_rq:/kbuild.1.se->load.weight.avg
176539 ± 22% +494.0% 1048576 sched_debug.cfs_rq:/kbuild.1.se->load.weight.max
1173 ± 53% +89267.3% 1048576 sched_debug.cfs_rq:/kbuild.1.se->load.weight.min
41159 ± 21% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.se->load.weight.stddev
44377 +1155.3% 557089 sched_debug.cfs_rq:/kbuild.1.se->sum_exec_runtime.avg
52129 +968.7% 557092 sched_debug.cfs_rq:/kbuild.1.se->sum_exec_runtime.max
36817 ± 2% +1413.1% 557086 sched_debug.cfs_rq:/kbuild.1.se->sum_exec_runtime.min
3173 ± 5% -99.9% 3.27 ±100% sched_debug.cfs_rq:/kbuild.1.se->sum_exec_runtime.stddev
2490096 -76.7% 580821 sched_debug.cfs_rq:/kbuild.1.se->vruntime.avg
2932328 -80.2% 581126 sched_debug.cfs_rq:/kbuild.1.se->vruntime.max
2062018 ± 2% -71.8% 580516 sched_debug.cfs_rq:/kbuild.1.se->vruntime.min
182836 ± 7% -99.8% 305.00 ±101% sched_debug.cfs_rq:/kbuild.1.se->vruntime.stddev
-2388892 -78.0% -525720 sched_debug.cfs_rq:/kbuild.1.spread0.avg
-2376905 -77.9% -525711 sched_debug.cfs_rq:/kbuild.1.spread0.max
-2400349 -78.1% -525729 sched_debug.cfs_rq:/kbuild.1.spread0.min
5008 ± 7% -99.8% 9.10 ±100% sched_debug.cfs_rq:/kbuild.1.spread0.stddev
59728 ± 4% -16.5% 49878 sched_debug.cfs_rq:/kbuild.1.tg_load_avg.max
2203 ± 20% -99.9% 1.82 ±223% sched_debug.cfs_rq:/kbuild.1.tg_load_avg.stddev
661.43 ± 3% +7393.3% 49562 sched_debug.cfs_rq:/kbuild.1.tg_load_avg_contrib.avg
2655 ± 11% +1778.5% 49878 sched_debug.cfs_rq:/kbuild.1.tg_load_avg_contrib.max
0.03 ±223% +1.8e+08% 49246 ± 2% sched_debug.cfs_rq:/kbuild.1.tg_load_avg_contrib.min
412.06 +140.8% 992.05 sched_debug.cfs_rq:/kbuild.1.util_avg.avg
0.03 ±223% +3.5e+06% 980.41 ± 2% sched_debug.cfs_rq:/kbuild.1.util_avg.min
368.29 -96.8% 11.64 ±100% sched_debug.cfs_rq:/kbuild.1.util_avg.stddev
0.00 ± 27% +7.1e+08% 6.35 ± 16% sched_debug.cfs_rq:/system.slice.MIN_vruntime.avg
0.00 ± 27% +2.9e+09% 25.72 sched_debug.cfs_rq:/system.slice.MIN_vruntime.max
0.00 ± 27% -70.0% 0.00 ± 25% sched_debug.cfs_rq:/system.slice.MIN_vruntime.min
739.56 ± 54% -90.0% 73.75 ± 36% sched_debug.cfs_rq:/system.slice.load_avg.max
183.19 ± 53% -82.7% 31.64 ± 37% sched_debug.cfs_rq:/system.slice.load_avg.stddev
0.00 ± 27% +7.3e+08% 6.45 ± 17% sched_debug.cfs_rq:/system.slice.max_vruntime.avg
0.00 ± 27% +2.9e+09% 26.12 sched_debug.cfs_rq:/system.slice.max_vruntime.max
0.00 ± 27% -70.0% 0.00 ± 25% sched_debug.cfs_rq:/system.slice.max_vruntime.min
30.65 ± 43% +4649.8% 1455 ± 30% sched_debug.cfs_rq:/system.slice.min_vruntime.avg
97.37 ± 36% +1415.3% 1475 ± 30% sched_debug.cfs_rq:/system.slice.min_vruntime.max
4.88 ±263% +29568.6% 1449 ± 30% sched_debug.cfs_rq:/system.slice.min_vruntime.min
739.56 ± 54% -90.0% 73.75 ± 36% sched_debug.cfs_rq:/system.slice.runnable_avg.max
183.44 ± 53% -82.8% 31.64 ± 37% sched_debug.cfs_rq:/system.slice.runnable_avg.stddev
698.25 ± 54% -92.7% 51.19 sched_debug.cfs_rq:/system.slice.se->avg.load_avg.max
173.10 ± 53% -87.3% 21.91 ± 5% sched_debug.cfs_rq:/system.slice.se->avg.load_avg.stddev
739.56 ± 54% -90.0% 73.76 ± 36% sched_debug.cfs_rq:/system.slice.se->avg.runnable_avg.max
183.45 ± 53% -82.8% 31.64 ± 37% sched_debug.cfs_rq:/system.slice.se->avg.runnable_avg.stddev
50.39 ± 53% -89.1% 5.49 ± 40% sched_debug.cfs_rq:/system.slice.se->avg.util_avg.avg
739.39 ± 54% -97.0% 22.29 ± 38% sched_debug.cfs_rq:/system.slice.se->avg.util_avg.max
183.41 ± 53% -94.8% 9.54 ± 39% sched_debug.cfs_rq:/system.slice.se->avg.util_avg.stddev
10507 ± 54% -92.1% 828.57 ± 5% sched_debug.cfs_rq:/system.slice.se->exec_start.stddev
796179 ± 45% -80.2% 157286 ± 19% sched_debug.cfs_rq:/system.slice.se->load.weight.max
2018 ± 70% +7694.2% 157286 ± 19% sched_debug.cfs_rq:/system.slice.se->load.weight.min
299826 ± 43% -100.0% 0.00 sched_debug.cfs_rq:/system.slice.se->load.weight.stddev
14.57 ±103% +10710.8% 1574 ± 30% sched_debug.cfs_rq:/system.slice.se->sum_exec_runtime.avg
98.32 ± 35% +1522.3% 1595 ± 29% sched_debug.cfs_rq:/system.slice.se->sum_exec_runtime.max
5.78 ±222% +27036.3% 1568 ± 30% sched_debug.cfs_rq:/system.slice.se->sum_exec_runtime.min
10560 ± 60% -95.1% 519.51 ± 12% sched_debug.cfs_rq:/system.slice.se->vruntime.stddev
781.68 ± 54% -89.7% 80.14 ± 30% sched_debug.cfs_rq:/system.slice.tg_load_avg.avg
783.75 ± 54% -88.9% 86.65 ± 31% sched_debug.cfs_rq:/system.slice.tg_load_avg.max
779.86 ± 54% -90.5% 73.87 ± 36% sched_debug.cfs_rq:/system.slice.tg_load_avg.min
739.36 ± 54% -90.0% 73.87 ± 36% sched_debug.cfs_rq:/system.slice.tg_load_avg_contrib.max
183.14 ± 53% -82.7% 31.69 ± 37% sched_debug.cfs_rq:/system.slice.tg_load_avg_contrib.stddev
50.40 ± 53% -89.1% 5.49 ± 40% sched_debug.cfs_rq:/system.slice.util_avg.avg
739.39 ± 54% -97.0% 22.29 ± 38% sched_debug.cfs_rq:/system.slice.util_avg.max
183.40 ± 53% -94.8% 9.54 ± 39% sched_debug.cfs_rq:/system.slice.util_avg.stddev
130276 ± 3% -53.9% 60023 ± 30% sched_debug.cpu.avg_idle.stddev
190561 +223.2% 615836 sched_debug.cpu.clock.avg
190566 +223.2% 615963 sched_debug.cpu.clock.max
190557 +223.1% 615766 sched_debug.cpu.clock.min
2.49 ± 2% +3195.3% 82.17 ± 10% sched_debug.cpu.clock.stddev
188349 +223.1% 608620 sched_debug.cpu.clock_task.avg
188650 +224.3% 611860 sched_debug.cpu.clock_task.max
186038 +224.3% 603373 sched_debug.cpu.clock_task.min
370.07 ± 40% +680.3% 2887 ± 40% sched_debug.cpu.clock_task.stddev
15323 ± 17% -97.1% 451.92 ± 5% sched_debug.cpu.curr->pid.avg
16512 ± 12% -74.5% 4215 ± 5% sched_debug.cpu.curr->pid.stddev
0.00 ± 5% +351.8% 0.00 ± 10% sched_debug.cpu.next_balance.stddev
0.38 +40.2% 0.54 sched_debug.cpu.nr_running.avg
1.11 ± 7% +4138.2% 47.09 sched_debug.cpu.nr_running.max
0.42 +1084.1% 4.99 sched_debug.cpu.nr_running.stddev
34116 -48.2% 17664 sched_debug.cpu.nr_switches.avg
66559 ± 6% +1204.5% 868266 sched_debug.cpu.nr_switches.max
26193 ± 2% -96.9% 813.00 ± 3% sched_debug.cpu.nr_switches.min
7164 ± 3% +1254.8% 97055 sched_debug.cpu.nr_switches.stddev
190557 +223.1% 615766 sched_debug.cpu_clk
189985 +223.8% 615197 sched_debug.ktime
193775 +219.7% 619543 sched_debug.sched_clk
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.
On 5/13/22 02:33, kernel test robot wrote: > > Greeting, > > FYI, we noticed a 2583.0% regression of kbuild.buildtime_per_iteration due to commit: > > > commit: 7aaa45f8374887d98cd479a6bdd3b4c74e6a981a ("[PATCH] cgroup/cpuset: Remove redundant cpu/node masks setup in cpuset_init_smp()") > url: https://github.com/intel-lab-lkp/linux/commits/Waiman-Long/cgroup-cpuset-Remove-redundant-cpu-node-masks-setup-in-cpuset_init_smp/20220425-101126 > base: https://git.kernel.org/cgit/linux/kernel/git/tj/cgroup.git for-next > patch link: https://lore.kernel.org/linux-mm/20220425020926.1264611-1-longman@redhat.com That link is the original v1 patch which is not correct. Is this the one that is being tested here? I have sent out another patch which shouldn't cause this kind of problem. Thanks, Longman > > in testcase: kbuild > on test machine: 88 threads 2 sockets Intel(R) Xeon(R) Gold 6238M CPU @ 2.10GHz with 128G memory > with following parameters: > > memory.high: 90% > memory.low: 50% > memory.max: max > pids.max: 10000 > runtime: 300s > nr_task: 50% > target: autoksyms_recursive > cpufreq_governor: performance > ucode: 0x500320a > > > > > If you fix the issue, kindly add following tag > Reported-by: kernel test robot <oliver.sang@intel.com> > > > Details are as below: > --------------------------------------------------------------------------------------------------> > > > To reproduce: > > git clone https://github.com/intel/lkp-tests.git > cd lkp-tests > sudo bin/lkp install job.yaml # job file is attached in this email > bin/lkp split-job --compatible job.yaml # generate the yaml file for lkp run > sudo bin/lkp run generated-yaml-file > > # if come across any failure that blocks the test, > # please remove ~/.lkp and /lkp dir to run from a clean state. > > ========================================================================================= > compiler/cpufreq_governor/kconfig/memory.high/memory.low/memory.max/nr_task/pids.max/rootfs/runtime/target/tbox_group/testcase/ucode: > gcc-11/performance/x86_64-rhel-8.3/90%/50%/max/50%/10000/debian-10.4-x86_64-20200603.cgz/300s/autoksyms_recursive/lkp-csl-2sp9/kbuild/0x500320a > > commit: > 4ab93063c8 ("cgroup: Add test_cpucg_weight_underprovisioned() testcase") > 7aaa45f837 ("cgroup/cpuset: Remove redundant cpu/node masks setup in cpuset_init_smp()") > > 4ab93063c83a2478 7aaa45f8374887d98cd479a6bdd > ---------------- --------------------------- > %stddev %change %stddev > \ | \ > 44.43 +2583.0% 1192 kbuild.buildtime_per_iteration > 311.14 +283.0% 1191 kbuild.time.elapsed_time > 311.14 +283.0% 1191 kbuild.time.elapsed_time.max > 363486 -4.7% 346532 kbuild.time.involuntary_context_switches > 7528 ± 2% -99.1% 68.00 kbuild.time.major_page_faults > 3.854e+08 -85.7% 55047165 kbuild.time.minor_page_faults > 2608 -96.4% 95.00 kbuild.time.percent_of_cpu_this_job_got > 1361 -89.5% 142.46 kbuild.time.system_time > 6756 -85.3% 991.32 kbuild.time.user_time > 1763962 -84.6% 270814 kbuild.time.voluntary_context_switches > 7.00 -85.7% 1.00 kbuild.workload > 1.886e+10 +439.9% 1.018e+11 cpuidle..time > 40524113 +397.6% 2.016e+08 ± 9% cpuidle..usage > 352.84 +249.7% 1233 uptime.boot > 22088 +378.4% 105677 uptime.idle > 68.86 +28.8 97.63 mpstat.cpu.all.idle% > 0.16 ± 3% -0.1 0.08 ± 8% mpstat.cpu.all.soft% > 5.23 -5.1 0.15 mpstat.cpu.all.sys% > 24.48 -23.5 0.96 mpstat.cpu.all.usr% > 68.83 +42.1% 97.83 vmstat.cpu.id > 24.00 -100.0% 0.00 vmstat.cpu.us > 26.00 +73.1% 45.00 vmstat.procs.r > 17438 -88.3% 2045 vmstat.system.cs > 1.749e+08 -70.6% 51378800 numa-numastat.node0.local_node > 1.75e+08 -70.6% 51378813 numa-numastat.node0.numa_hit > 50974 ± 50% -100.0% 13.50 ± 75% numa-numastat.node0.other_node > 1.75e+08 -100.0% 82.83 ±223% numa-numastat.node1.local_node > 1.75e+08 -100.0% 79394 numa-numastat.node1.numa_hit > 28316 ± 91% +180.1% 79311 numa-numastat.node1.other_node > 6700 ± 4% +69.9% 11385 ± 3% meminfo.Active > 6540 ± 4% +69.1% 11063 meminfo.Active(anon) > 223421 +31.7% 294262 ± 2% meminfo.AnonHugePages > 1300267 +31.4% 1708081 meminfo.AnonPages > 1576268 +25.7% 1981479 meminfo.Committed_AS > 1338713 +28.5% 1719792 meminfo.Inactive > 1338565 +28.4% 1719004 meminfo.Inactive(anon) > 18332 -11.0% 16310 meminfo.KernelStack > 93129 -43.8% 52361 meminfo.Mapped > 14656 ± 2% -8.9% 13355 meminfo.PageTables > 43576 ± 3% -52.0% 20896 meminfo.Shmem > 7581989 -10.5% 6785067 meminfo.max_used_kB > 875.33 -100.0% 0.00 turbostat.Avg_MHz > 31.74 -31.7 0.00 turbostat.Busy% > 2764 -100.0% 0.00 turbostat.Bzy_MHz > 173892 ± 10% -100.0% 0.00 turbostat.C1 > 0.02 ± 17% -0.0 0.00 turbostat.C1% > 33508125 ± 20% -100.0% 0.00 turbostat.C1E > 48.18 ± 44% -48.2 0.00 turbostat.C1E% > 6776884 ±102% -100.0% 0.00 turbostat.C6 > 20.30 ±104% -20.3 0.00 turbostat.C6% > 66.40 ± 4% -100.0% 0.00 turbostat.CPU%c1 > 1.86 ±139% -100.0% 0.00 turbostat.CPU%c6 > 59.17 -100.0% 0.00 turbostat.CoreTmp > 0.33 -100.0% 0.00 turbostat.IPC > 55232878 -100.0% 0.00 turbostat.IRQ > 56380 ± 6% -100.0% 0.00 turbostat.POLL > 59.00 -100.0% 0.00 turbostat.PkgTmp > 215.76 -100.0% 0.00 turbostat.PkgWatt > 16.16 -100.0% 0.00 turbostat.RAMWatt > 2095 -100.0% 0.00 turbostat.TSC_MHz > 813.00 ± 11% +237.1% 2740 numa-vmstat.node0.nr_active_anon > 183776 +131.6% 425651 numa-vmstat.node0.nr_anon_pages > 186417 +129.7% 428211 numa-vmstat.node0.nr_inactive_anon > 1978 +68.0% 3324 numa-vmstat.node0.nr_page_table_pages > 3282 ± 7% +53.1% 5023 numa-vmstat.node0.nr_shmem > 813.00 ± 11% +237.1% 2740 numa-vmstat.node0.nr_zone_active_anon > 186416 +129.7% 428211 numa-vmstat.node0.nr_zone_inactive_anon > 1.75e+08 -70.6% 51378302 numa-vmstat.node0.numa_hit > 1.749e+08 -70.6% 51378288 numa-vmstat.node0.numa_local > 50973 ± 50% -100.0% 13.50 ± 75% numa-vmstat.node0.numa_other > 823.33 ± 5% -97.5% 20.33 ± 12% numa-vmstat.node1.nr_active_anon > 139750 ± 2% -99.9% 88.17 ±114% numa-vmstat.node1.nr_anon_pages > 157716 ±183% -97.6% 3733 ±208% numa-vmstat.node1.nr_file_pages > 146649 ± 2% -99.8% 265.00 ± 56% numa-vmstat.node1.nr_inactive_anon > 8288 ± 4% -31.8% 5653 ± 5% numa-vmstat.node1.nr_kernel_stack > 8789 ± 47% -99.6% 32.00 ±120% numa-vmstat.node1.nr_mapped > 1658 -99.8% 4.00 ± 73% numa-vmstat.node1.nr_page_table_pages > 7610 ± 3% -97.4% 197.17 ± 27% numa-vmstat.node1.nr_shmem > 11225 ±101% -74.3% 2883 ± 9% numa-vmstat.node1.nr_slab_reclaimable > 20034 ± 3% -25.8% 14861 ± 4% numa-vmstat.node1.nr_slab_unreclaimable > 823.33 ± 5% -97.5% 20.33 ± 12% numa-vmstat.node1.nr_zone_active_anon > 146648 ± 2% -99.8% 265.00 ± 56% numa-vmstat.node1.nr_zone_inactive_anon > 1.75e+08 -100.0% 79394 numa-vmstat.node1.numa_hit > 1.75e+08 -100.0% 82.83 ±223% numa-vmstat.node1.numa_local > 28312 ± 91% +180.1% 79311 numa-vmstat.node1.numa_other > 1642 ± 4% +68.3% 2763 proc-vmstat.nr_active_anon > 322313 +32.5% 427139 proc-vmstat.nr_anon_pages > 864421 -1.5% 851593 proc-vmstat.nr_file_pages > 332108 +29.4% 429879 proc-vmstat.nr_inactive_anon > 18183 -10.3% 16310 proc-vmstat.nr_kernel_stack > 23431 -43.6% 13208 proc-vmstat.nr_mapped > 3550 ± 2% -6.0% 3337 proc-vmstat.nr_page_table_pages > 11164 ± 3% -53.2% 5223 proc-vmstat.nr_shmem > 46151 -2.2% 45116 proc-vmstat.nr_slab_reclaimable > 46770 -8.6% 42749 proc-vmstat.nr_slab_unreclaimable > 1642 ± 4% +68.3% 2763 proc-vmstat.nr_zone_active_anon > 332108 +29.4% 429879 proc-vmstat.nr_zone_inactive_anon > 360277 ± 30% -99.8% 680.17 ±112% proc-vmstat.numa_hint_faults > 357678 ± 31% -99.8% 680.17 ±112% proc-vmstat.numa_hint_faults_local > 3.5e+08 -85.3% 51462030 proc-vmstat.numa_hit > 3.499e+08 -85.3% 51382705 proc-vmstat.numa_local > 10372 ±129% -100.0% 0.00 proc-vmstat.numa_pages_migrated > 790289 ± 15% -75.5% 193877 ± 7% proc-vmstat.numa_pte_updates > 290045 -86.1% 40379 proc-vmstat.pgactivate > 3.498e+08 -85.3% 51462440 proc-vmstat.pgalloc_normal > 3.871e+08 -85.1% 57578373 proc-vmstat.pgfault > 3.497e+08 -85.3% 51397471 proc-vmstat.pgfree > 10372 ±129% -100.0% 0.00 proc-vmstat.pgmigrate_success > 2920438 -78.6% 624274 proc-vmstat.pgreuse > 18128 -85.7% 2594 proc-vmstat.thp_fault_alloc > 771909 -85.7% 110390 proc-vmstat.unevictable_pgs_culled > 3411 ± 11% +226.2% 11127 numa-meminfo.node0.Active > 3251 ± 11% +237.3% 10966 numa-meminfo.node0.Active(anon) > 188175 ± 6% +56.0% 293523 ± 2% numa-meminfo.node0.AnonHugePages > 737315 +130.8% 1701834 numa-meminfo.node0.AnonPages > 1657848 ± 5% +85.9% 3081876 ± 4% numa-meminfo.node0.AnonPages.max > 748023 +129.0% 1712859 numa-meminfo.node0.Inactive > 747875 +128.9% 1712108 numa-meminfo.node0.Inactive(anon) > 4306840 ± 28% +35.6% 5839514 numa-meminfo.node0.MemUsed > 7772 ± 3% +71.1% 13297 numa-meminfo.node0.PageTables > 13182 ± 8% +52.5% 20099 numa-meminfo.node0.Shmem > 3307 ± 5% -92.7% 242.00 ±147% numa-meminfo.node1.Active > 3307 ± 5% -97.5% 81.33 ± 12% numa-meminfo.node1.Active(anon) > 34918 ± 34% -100.0% 0.00 numa-meminfo.node1.AnonHugePages > 557333 ± 2% -99.9% 353.33 ±114% numa-meminfo.node1.AnonPages > 1876230 ± 3% -100.0% 496.67 ± 96% numa-meminfo.node1.AnonPages.max > 631836 ±183% -97.6% 14935 ±208% numa-meminfo.node1.FilePages > 586137 ± 2% -99.8% 1096 ± 51% numa-meminfo.node1.Inactive > 586137 ± 2% -99.8% 1061 ± 56% numa-meminfo.node1.Inactive(anon) > 44909 ±101% -74.3% 11533 ± 9% numa-meminfo.node1.KReclaimable > 8244 ± 4% -31.4% 5653 ± 5% numa-meminfo.node1.KernelStack > 35330 ± 46% -99.6% 129.17 ±119% numa-meminfo.node1.Mapped > 1612958 ± 74% -79.4% 333010 ± 7% numa-meminfo.node1.MemUsed > 6577 -99.8% 16.17 ± 73% numa-meminfo.node1.PageTables > 44909 ±101% -74.3% 11533 ± 9% numa-meminfo.node1.SReclaimable > 80135 ± 3% -25.8% 59446 ± 4% numa-meminfo.node1.SUnreclaim > 31315 ± 3% -97.5% 789.17 ± 27% numa-meminfo.node1.Shmem > 125044 ± 38% -43.2% 70980 ± 4% numa-meminfo.node1.Slab > 1.978e+10 -95.6% 8.797e+08 perf-stat.i.branch-instructions > 5.616e+08 -95.6% 24754884 ± 8% perf-stat.i.branch-misses > 1.495e+08 -94.7% 7868663 ± 5% perf-stat.i.cache-misses > 6.176e+08 -94.5% 34193232 ± 36% perf-stat.i.cache-references > 17499 -88.4% 2029 perf-stat.i.context-switches > 0.93 ± 3% +36.6% 1.27 ± 5% perf-stat.i.cpi > 7.611e+10 -93.3% 5.121e+09 ± 4% perf-stat.i.cpu-cycles > 1275 -100.0% 0.04 ± 28% perf-stat.i.cpu-migrations > 1029 ± 5% -31.3% 707.28 ± 2% perf-stat.i.cycles-between-cache-misses > 14172722 -94.1% 836284 ± 47% perf-stat.i.dTLB-load-misses > 2.503e+10 -95.5% 1.122e+09 perf-stat.i.dTLB-loads > 7917427 -95.8% 332811 ± 15% perf-stat.i.dTLB-store-misses > 1.244e+10 -95.4% 5.666e+08 perf-stat.i.dTLB-stores > 24631063 -92.3% 1894591 ± 2% perf-stat.i.iTLB-load-misses > 71323368 -94.9% 3665231 ± 6% perf-stat.i.iTLB-loads > 9.005e+10 -95.5% 4.054e+09 perf-stat.i.instructions > 3571 -39.7% 2152 perf-stat.i.instructions-per-iTLB-miss > 1.10 ± 2% -28.1% 0.79 ± 5% perf-stat.i.ipc > 24.38 ± 2% -99.5% 0.11 ± 23% perf-stat.i.major-faults > 0.86 -93.3% 0.06 ± 4% perf-stat.i.metric.GHz > 658.74 -95.6% 29.19 perf-stat.i.metric.M/sec > 1236978 -96.1% 48019 perf-stat.i.minor-faults > 24.93 ± 5% -18.7 6.26 ± 22% perf-stat.i.node-load-miss-rate% > 6765575 -97.4% 178075 ± 26% perf-stat.i.node-load-misses > 55797319 -94.4% 3120263 perf-stat.i.node-loads > 18.32 ± 4% -9.7 8.62 ± 22% perf-stat.i.node-store-miss-rate% > 1648658 -97.0% 50137 ± 26% perf-stat.i.node-store-misses > 11762943 -95.0% 585083 perf-stat.i.node-stores > 1237003 -96.1% 48019 perf-stat.i.page-faults > 6.86 +23.4% 8.46 ± 38% perf-stat.overall.MPKI > 0.85 +49.5% 1.26 ± 5% perf-stat.overall.cpi > 508.96 +27.9% 651.10 perf-stat.overall.cycles-between-cache-misses > 25.67 +8.5 34.14 ± 3% perf-stat.overall.iTLB-load-miss-rate% > 3655 -41.4% 2140 perf-stat.overall.instructions-per-iTLB-miss > 1.18 -32.9% 0.79 ± 5% perf-stat.overall.ipc > 10.82 -5.4 5.38 ± 24% perf-stat.overall.node-load-miss-rate% > 12.30 -4.4 7.85 ± 23% perf-stat.overall.node-store-miss-rate% > 4.002e+12 +20.6% 4.826e+12 perf-stat.overall.path-length > 1.972e+10 -95.5% 8.791e+08 perf-stat.ps.branch-instructions > 5.599e+08 -95.6% 24737459 ± 8% perf-stat.ps.branch-misses > 1.491e+08 -94.7% 7862895 ± 5% perf-stat.ps.cache-misses > 6.158e+08 -94.5% 34168138 ± 36% perf-stat.ps.cache-references > 17447 -88.4% 2027 perf-stat.ps.context-switches > 7.588e+10 -93.3% 5.118e+09 ± 4% perf-stat.ps.cpu-cycles > 1272 -100.0% 0.04 ± 28% perf-stat.ps.cpu-migrations > 14129808 -94.1% 835661 ± 47% perf-stat.ps.dTLB-load-misses > 2.495e+10 -95.5% 1.121e+09 perf-stat.ps.dTLB-loads > 7893270 -95.8% 332570 ± 15% perf-stat.ps.dTLB-store-misses > 1.24e+10 -95.4% 5.662e+08 perf-stat.ps.dTLB-stores > 24556703 -92.3% 1893249 ± 2% perf-stat.ps.iTLB-load-misses > 71106326 -94.8% 3662587 ± 6% perf-stat.ps.iTLB-loads > 8.978e+10 -95.5% 4.051e+09 perf-stat.ps.instructions > 24.32 ± 2% -99.5% 0.11 ± 24% perf-stat.ps.major-faults > 1233214 -96.1% 47986 perf-stat.ps.minor-faults > 6745825 -97.4% 177950 ± 26% perf-stat.ps.node-load-misses > 55627061 -94.4% 3117959 perf-stat.ps.node-loads > 1644068 -97.0% 50102 ± 26% perf-stat.ps.node-store-misses > 11727184 -95.0% 584651 perf-stat.ps.node-stores > 1233238 -96.1% 47986 perf-stat.ps.page-faults > 2.802e+13 -82.8% 4.826e+12 perf-stat.total.instructions > 0.39 +3586.3% 14.32 ± 4% sched_debug.cfs_rq:/.h_nr_running.avg > 1.11 ± 7% +4138.2% 47.09 sched_debug.cfs_rq:/.h_nr_running.max > 0.43 +4683.1% 20.72 sched_debug.cfs_rq:/.h_nr_running.stddev > 14081 ± 20% +2107.0% 310782 ± 4% sched_debug.cfs_rq:/.load.avg > 407170 ± 49% +170.4% 1101004 sched_debug.cfs_rq:/.load.max > 52684 ± 42% +768.5% 457567 sched_debug.cfs_rq:/.load.stddev > 37.01 ± 46% +723.1% 304.63 ± 4% sched_debug.cfs_rq:/.load_avg.avg > 144.31 ± 45% +210.7% 448.32 sched_debug.cfs_rq:/.load_avg.stddev > 2488735 -92.5% 187680 ± 3% sched_debug.cfs_rq:/.min_vruntime.avg > 2938975 -80.2% 581666 sched_debug.cfs_rq:/.min_vruntime.max > 2060745 ± 2% -99.5% 10250 ± 34% sched_debug.cfs_rq:/.min_vruntime.min > 182406 ± 6% +39.2% 253824 sched_debug.cfs_rq:/.min_vruntime.stddev > 0.39 -23.7% 0.30 ± 4% sched_debug.cfs_rq:/.nr_running.avg > 400.60 +3272.5% 13510 ± 5% sched_debug.cfs_rq:/.runnable_avg.avg > 971.78 ± 2% +4430.1% 44022 sched_debug.cfs_rq:/.runnable_avg.max > 370.10 +5173.5% 19517 sched_debug.cfs_rq:/.runnable_avg.stddev > 12579 ±1058% -3228.1% -393486 sched_debug.cfs_rq:/.spread0.avg > 462824 ± 24% -99.9% 540.01 ± 92% sched_debug.cfs_rq:/.spread0.max > 182401 ± 6% +39.2% 253852 sched_debug.cfs_rq:/.spread0.stddev > 400.13 -25.0% 300.07 ± 4% sched_debug.cfs_rq:/.util_avg.avg > 369.97 +18.4% 437.88 sched_debug.cfs_rq:/.util_avg.stddev > 10.90 ± 10% -87.5% 1.37 ± 48% sched_debug.cfs_rq:/.util_est_enqueued.avg > 454.44 ± 14% -90.3% 44.23 ± 3% sched_debug.cfs_rq:/.util_est_enqueued.max > 60.82 ± 13% -90.9% 5.53 ± 14% sched_debug.cfs_rq:/.util_est_enqueued.stddev > 0.92 ± 36% -90.0% 0.09 ±176% sched_debug.cfs_rq:/init.scope.load_avg.avg > 1.08 ± 17% -91.5% 0.09 ±176% sched_debug.cfs_rq:/init.scope.load_avg.max > 0.50 ±107% +2506.1% 12.97 ± 57% sched_debug.cfs_rq:/init.scope.min_vruntime.avg > 0.54 ±104% +2288.9% 13.00 ± 57% sched_debug.cfs_rq:/init.scope.min_vruntime.max > 0.45 ±111% +2768.6% 12.93 ± 57% sched_debug.cfs_rq:/init.scope.min_vruntime.min > 0.92 ± 36% -90.0% 0.09 ±176% sched_debug.cfs_rq:/init.scope.runnable_avg.avg > 1.08 ± 17% -91.5% 0.09 ±176% sched_debug.cfs_rq:/init.scope.runnable_avg.max > 0.51 ±123% +1809.6% 9.78 ± 40% sched_debug.cfs_rq:/init.scope.se->sum_exec_runtime.avg > 0.52 ±122% +1769.0% 9.78 ± 40% sched_debug.cfs_rq:/init.scope.se->sum_exec_runtime.max > 0.50 ±126% +1852.0% 9.77 ± 40% sched_debug.cfs_rq:/init.scope.se->sum_exec_runtime.min > 927421 ±116% -88.6% 105436 ± 29% sched_debug.cfs_rq:/init.scope.se->vruntime.avg > 927672 ±116% -88.6% 105945 ± 29% sched_debug.cfs_rq:/init.scope.se->vruntime.max > 927171 ±117% -88.7% 104926 ± 29% sched_debug.cfs_rq:/init.scope.se->vruntime.min > -903221 -88.3% -106040 sched_debug.cfs_rq:/init.scope.spread0.avg > -903221 -88.3% -106040 sched_debug.cfs_rq:/init.scope.spread0.max > -903221 -88.3% -106040 sched_debug.cfs_rq:/init.scope.spread0.min > 1.08 ± 17% -91.5% 0.09 ±176% sched_debug.cfs_rq:/init.scope.tg_load_avg.avg > 1.08 ± 17% -91.5% 0.09 ±176% sched_debug.cfs_rq:/init.scope.tg_load_avg.max > 1.08 ± 17% -91.5% 0.09 ±176% sched_debug.cfs_rq:/init.scope.tg_load_avg.min > 0.92 ± 36% -90.0% 0.09 ±176% sched_debug.cfs_rq:/init.scope.tg_load_avg_contrib.avg > 1.08 ± 17% -91.5% 0.09 ±176% sched_debug.cfs_rq:/init.scope.tg_load_avg_contrib.max > 0.92 ± 36% -96.4% 0.03 ±111% sched_debug.cfs_rq:/init.scope.util_avg.avg > 1.08 ± 17% -96.9% 0.03 ±111% sched_debug.cfs_rq:/init.scope.util_avg.max > 109.53 ±107% +50485.6% 55405 sched_debug.cfs_rq:/kbuild.1.MIN_vruntime.avg > 9638 ±107% +475.0% 55414 sched_debug.cfs_rq:/kbuild.1.MIN_vruntime.max > -0.30 -1.9e+07% 55396 sched_debug.cfs_rq:/kbuild.1.MIN_vruntime.min > 0.39 ± 2% +11812.2% 46.70 sched_debug.cfs_rq:/kbuild.1.h_nr_running.avg > 1.00 +4594.2% 46.94 sched_debug.cfs_rq:/kbuild.1.h_nr_running.max > 410733 ± 2% +11823.3% 48972868 sched_debug.cfs_rq:/kbuild.1.load.avg > 1048576 +4594.2% 49221905 sched_debug.cfs_rq:/kbuild.1.load.max > 659.13 ± 3% +7394.5% 49398 sched_debug.cfs_rq:/kbuild.1.load_avg.avg > 2647 ± 11% +1777.5% 49713 sched_debug.cfs_rq:/kbuild.1.load_avg.max > 0.03 ±223% +1.8e+08% 49083 ± 2% sched_debug.cfs_rq:/kbuild.1.load_avg.min > 109.53 ±107% +50489.2% 55409 sched_debug.cfs_rq:/kbuild.1.max_vruntime.avg > 9638 ±107% +475.0% 55418 sched_debug.cfs_rq:/kbuild.1.max_vruntime.max > -0.30 -1.9e+07% 55400 sched_debug.cfs_rq:/kbuild.1.max_vruntime.min > 87264 -36.5% 55405 sched_debug.cfs_rq:/kbuild.1.min_vruntime.avg > 99270 -44.2% 55414 sched_debug.cfs_rq:/kbuild.1.min_vruntime.max > 75767 -26.9% 55396 sched_debug.cfs_rq:/kbuild.1.min_vruntime.min > 5024 ± 7% -99.8% 9.08 ±100% sched_debug.cfs_rq:/kbuild.1.min_vruntime.stddev > 0.39 ± 2% +11812.2% 46.70 sched_debug.cfs_rq:/kbuild.1.nr_running.avg > 1.00 +4594.2% 46.94 sched_debug.cfs_rq:/kbuild.1.nr_running.max > 35.29 ± 22% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.load_avg.avg > 919.11 ± 14% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.load_avg.max > 156.52 ± 17% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.load_avg.stddev > 14.27 ± 27% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.runnable_avg.avg > 419.11 ± 25% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.runnable_avg.max > 65.99 ± 27% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.runnable_avg.stddev > 14.27 ± 27% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.util_avg.avg > 419.11 ± 25% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.util_avg.max > 65.99 ± 27% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.removed.util_avg.stddev > 412.27 +10527.3% 43813 sched_debug.cfs_rq:/kbuild.1.runnable_avg.avg > 967.14 ± 2% +4443.7% 43944 sched_debug.cfs_rq:/kbuild.1.runnable_avg.max > 0.03 ±223% +1.6e+08% 43681 sched_debug.cfs_rq:/kbuild.1.runnable_avg.min > 368.39 -64.4% 131.08 ±101% sched_debug.cfs_rq:/kbuild.1.runnable_avg.stddev > 13.05 ± 10% +7646.3% 1011 sched_debug.cfs_rq:/kbuild.1.se->avg.load_avg.avg > 87.56 ± 16% +1069.5% 1023 sched_debug.cfs_rq:/kbuild.1.se->avg.load_avg.max > 412.13 +10524.7% 43787 sched_debug.cfs_rq:/kbuild.1.se->avg.runnable_avg.avg > 967.47 ± 2% +4439.4% 43917 sched_debug.cfs_rq:/kbuild.1.se->avg.runnable_avg.max > 0.03 ±223% +1.6e+08% 43657 sched_debug.cfs_rq:/kbuild.1.se->avg.runnable_avg.min > 368.51 -64.7% 129.95 ±101% sched_debug.cfs_rq:/kbuild.1.se->avg.runnable_avg.stddev > 411.92 +141.6% 995.10 sched_debug.cfs_rq:/kbuild.1.se->avg.util_avg.avg > 0.03 ±223% +3.5e+06% 983.40 ± 2% sched_debug.cfs_rq:/kbuild.1.se->avg.util_avg.min > 368.41 -96.8% 11.70 ±100% sched_debug.cfs_rq:/kbuild.1.se->avg.util_avg.stddev > 188343 +220.5% 603616 sched_debug.cfs_rq:/kbuild.1.se->exec_start.avg > 188644 +220.2% 604099 sched_debug.cfs_rq:/kbuild.1.se->exec_start.max > 186221 +223.9% 603133 sched_debug.cfs_rq:/kbuild.1.se->exec_start.min > 51889 ± 7% +1920.8% 1048576 sched_debug.cfs_rq:/kbuild.1.se->load.weight.avg > 176539 ± 22% +494.0% 1048576 sched_debug.cfs_rq:/kbuild.1.se->load.weight.max > 1173 ± 53% +89267.3% 1048576 sched_debug.cfs_rq:/kbuild.1.se->load.weight.min > 41159 ± 21% -100.0% 0.00 sched_debug.cfs_rq:/kbuild.1.se->load.weight.stddev > 44377 +1155.3% 557089 sched_debug.cfs_rq:/kbuild.1.se->sum_exec_runtime.avg > 52129 +968.7% 557092 sched_debug.cfs_rq:/kbuild.1.se->sum_exec_runtime.max > 36817 ± 2% +1413.1% 557086 sched_debug.cfs_rq:/kbuild.1.se->sum_exec_runtime.min > 3173 ± 5% -99.9% 3.27 ±100% sched_debug.cfs_rq:/kbuild.1.se->sum_exec_runtime.stddev > 2490096 -76.7% 580821 sched_debug.cfs_rq:/kbuild.1.se->vruntime.avg > 2932328 -80.2% 581126 sched_debug.cfs_rq:/kbuild.1.se->vruntime.max > 2062018 ± 2% -71.8% 580516 sched_debug.cfs_rq:/kbuild.1.se->vruntime.min > 182836 ± 7% -99.8% 305.00 ±101% sched_debug.cfs_rq:/kbuild.1.se->vruntime.stddev > -2388892 -78.0% -525720 sched_debug.cfs_rq:/kbuild.1.spread0.avg > -2376905 -77.9% -525711 sched_debug.cfs_rq:/kbuild.1.spread0.max > -2400349 -78.1% -525729 sched_debug.cfs_rq:/kbuild.1.spread0.min > 5008 ± 7% -99.8% 9.10 ±100% sched_debug.cfs_rq:/kbuild.1.spread0.stddev > 59728 ± 4% -16.5% 49878 sched_debug.cfs_rq:/kbuild.1.tg_load_avg.max > 2203 ± 20% -99.9% 1.82 ±223% sched_debug.cfs_rq:/kbuild.1.tg_load_avg.stddev > 661.43 ± 3% +7393.3% 49562 sched_debug.cfs_rq:/kbuild.1.tg_load_avg_contrib.avg > 2655 ± 11% +1778.5% 49878 sched_debug.cfs_rq:/kbuild.1.tg_load_avg_contrib.max > 0.03 ±223% +1.8e+08% 49246 ± 2% sched_debug.cfs_rq:/kbuild.1.tg_load_avg_contrib.min > 412.06 +140.8% 992.05 sched_debug.cfs_rq:/kbuild.1.util_avg.avg > 0.03 ±223% +3.5e+06% 980.41 ± 2% sched_debug.cfs_rq:/kbuild.1.util_avg.min > 368.29 -96.8% 11.64 ±100% sched_debug.cfs_rq:/kbuild.1.util_avg.stddev > 0.00 ± 27% +7.1e+08% 6.35 ± 16% sched_debug.cfs_rq:/system.slice.MIN_vruntime.avg > 0.00 ± 27% +2.9e+09% 25.72 sched_debug.cfs_rq:/system.slice.MIN_vruntime.max > 0.00 ± 27% -70.0% 0.00 ± 25% sched_debug.cfs_rq:/system.slice.MIN_vruntime.min > 739.56 ± 54% -90.0% 73.75 ± 36% sched_debug.cfs_rq:/system.slice.load_avg.max > 183.19 ± 53% -82.7% 31.64 ± 37% sched_debug.cfs_rq:/system.slice.load_avg.stddev > 0.00 ± 27% +7.3e+08% 6.45 ± 17% sched_debug.cfs_rq:/system.slice.max_vruntime.avg > 0.00 ± 27% +2.9e+09% 26.12 sched_debug.cfs_rq:/system.slice.max_vruntime.max > 0.00 ± 27% -70.0% 0.00 ± 25% sched_debug.cfs_rq:/system.slice.max_vruntime.min > 30.65 ± 43% +4649.8% 1455 ± 30% sched_debug.cfs_rq:/system.slice.min_vruntime.avg > 97.37 ± 36% +1415.3% 1475 ± 30% sched_debug.cfs_rq:/system.slice.min_vruntime.max > 4.88 ±263% +29568.6% 1449 ± 30% sched_debug.cfs_rq:/system.slice.min_vruntime.min > 739.56 ± 54% -90.0% 73.75 ± 36% sched_debug.cfs_rq:/system.slice.runnable_avg.max > 183.44 ± 53% -82.8% 31.64 ± 37% sched_debug.cfs_rq:/system.slice.runnable_avg.stddev > 698.25 ± 54% -92.7% 51.19 sched_debug.cfs_rq:/system.slice.se->avg.load_avg.max > 173.10 ± 53% -87.3% 21.91 ± 5% sched_debug.cfs_rq:/system.slice.se->avg.load_avg.stddev > 739.56 ± 54% -90.0% 73.76 ± 36% sched_debug.cfs_rq:/system.slice.se->avg.runnable_avg.max > 183.45 ± 53% -82.8% 31.64 ± 37% sched_debug.cfs_rq:/system.slice.se->avg.runnable_avg.stddev > 50.39 ± 53% -89.1% 5.49 ± 40% sched_debug.cfs_rq:/system.slice.se->avg.util_avg.avg > 739.39 ± 54% -97.0% 22.29 ± 38% sched_debug.cfs_rq:/system.slice.se->avg.util_avg.max > 183.41 ± 53% -94.8% 9.54 ± 39% sched_debug.cfs_rq:/system.slice.se->avg.util_avg.stddev > 10507 ± 54% -92.1% 828.57 ± 5% sched_debug.cfs_rq:/system.slice.se->exec_start.stddev > 796179 ± 45% -80.2% 157286 ± 19% sched_debug.cfs_rq:/system.slice.se->load.weight.max > 2018 ± 70% +7694.2% 157286 ± 19% sched_debug.cfs_rq:/system.slice.se->load.weight.min > 299826 ± 43% -100.0% 0.00 sched_debug.cfs_rq:/system.slice.se->load.weight.stddev > 14.57 ±103% +10710.8% 1574 ± 30% sched_debug.cfs_rq:/system.slice.se->sum_exec_runtime.avg > 98.32 ± 35% +1522.3% 1595 ± 29% sched_debug.cfs_rq:/system.slice.se->sum_exec_runtime.max > 5.78 ±222% +27036.3% 1568 ± 30% sched_debug.cfs_rq:/system.slice.se->sum_exec_runtime.min > 10560 ± 60% -95.1% 519.51 ± 12% sched_debug.cfs_rq:/system.slice.se->vruntime.stddev > 781.68 ± 54% -89.7% 80.14 ± 30% sched_debug.cfs_rq:/system.slice.tg_load_avg.avg > 783.75 ± 54% -88.9% 86.65 ± 31% sched_debug.cfs_rq:/system.slice.tg_load_avg.max > 779.86 ± 54% -90.5% 73.87 ± 36% sched_debug.cfs_rq:/system.slice.tg_load_avg.min > 739.36 ± 54% -90.0% 73.87 ± 36% sched_debug.cfs_rq:/system.slice.tg_load_avg_contrib.max > 183.14 ± 53% -82.7% 31.69 ± 37% sched_debug.cfs_rq:/system.slice.tg_load_avg_contrib.stddev > 50.40 ± 53% -89.1% 5.49 ± 40% sched_debug.cfs_rq:/system.slice.util_avg.avg > 739.39 ± 54% -97.0% 22.29 ± 38% sched_debug.cfs_rq:/system.slice.util_avg.max > 183.40 ± 53% -94.8% 9.54 ± 39% sched_debug.cfs_rq:/system.slice.util_avg.stddev > 130276 ± 3% -53.9% 60023 ± 30% sched_debug.cpu.avg_idle.stddev > 190561 +223.2% 615836 sched_debug.cpu.clock.avg > 190566 +223.2% 615963 sched_debug.cpu.clock.max > 190557 +223.1% 615766 sched_debug.cpu.clock.min > 2.49 ± 2% +3195.3% 82.17 ± 10% sched_debug.cpu.clock.stddev > 188349 +223.1% 608620 sched_debug.cpu.clock_task.avg > 188650 +224.3% 611860 sched_debug.cpu.clock_task.max > 186038 +224.3% 603373 sched_debug.cpu.clock_task.min > 370.07 ± 40% +680.3% 2887 ± 40% sched_debug.cpu.clock_task.stddev > 15323 ± 17% -97.1% 451.92 ± 5% sched_debug.cpu.curr->pid.avg > 16512 ± 12% -74.5% 4215 ± 5% sched_debug.cpu.curr->pid.stddev > 0.00 ± 5% +351.8% 0.00 ± 10% sched_debug.cpu.next_balance.stddev > 0.38 +40.2% 0.54 sched_debug.cpu.nr_running.avg > 1.11 ± 7% +4138.2% 47.09 sched_debug.cpu.nr_running.max > 0.42 +1084.1% 4.99 sched_debug.cpu.nr_running.stddev > 34116 -48.2% 17664 sched_debug.cpu.nr_switches.avg > 66559 ± 6% +1204.5% 868266 sched_debug.cpu.nr_switches.max > 26193 ± 2% -96.9% 813.00 ± 3% sched_debug.cpu.nr_switches.min > 7164 ± 3% +1254.8% 97055 sched_debug.cpu.nr_switches.stddev > 190557 +223.1% 615766 sched_debug.cpu_clk > 189985 +223.8% 615197 sched_debug.ktime > 193775 +219.7% 619543 sched_debug.sched_clk > > > > > 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. > >
diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c index 9390bfd9f1cd..a2e15a43397e 100644 --- a/kernel/cgroup/cpuset.c +++ b/kernel/cgroup/cpuset.c @@ -2961,6 +2961,9 @@ static void cpuset_bind(struct cgroup_subsys_state *root_css) percpu_down_write(&cpuset_rwsem); spin_lock_irq(&callback_lock); + cpumask_copy(top_cpuset.effective_cpus, cpu_active_mask); + top_cpuset.effective_mems = node_states[N_MEMORY]; + if (is_in_v2_mode()) { cpumask_copy(top_cpuset.cpus_allowed, cpu_possible_mask); top_cpuset.mems_allowed = node_possible_map; @@ -3390,13 +3393,6 @@ static struct notifier_block cpuset_track_online_nodes_nb = { */ void __init cpuset_init_smp(void) { - cpumask_copy(top_cpuset.cpus_allowed, cpu_active_mask); - top_cpuset.mems_allowed = node_states[N_MEMORY]; - top_cpuset.old_mems_allowed = top_cpuset.mems_allowed; - - cpumask_copy(top_cpuset.effective_cpus, cpu_active_mask); - top_cpuset.effective_mems = node_states[N_MEMORY]; - register_hotmemory_notifier(&cpuset_track_online_nodes_nb); cpuset_migrate_mm_wq = alloc_ordered_workqueue("cpuset_migrate_mm", 0);
There are 3 places where the cpu and node masks of the top cpuset can be initialized in the order they are executed: 1) start_kernel -> cpuset_init() 2) start_kernel -> cgroup_init() -> cpuset_bind() 3) kernel_init_freeable() -> do_basic_setup() -> cpuset_init_smp() The first cpuset_init() function just sets all the bits in the masks. The last one executed is cpuset_init_smp() which sets up cpu and node masks suitable for v1, but not v2. cpuset_bind() does the right setup for both v1 and v2 assuming that effective_mems and effective_cpus have been set up properly which is not strictly the case here. As a result, cpu and memory node hot add may fail to update the cpu and node masks of the top cpuset to include the newly added cpu or node in a cgroup v2 environment. To fix this problem, the redundant cpus_allowed and mems_allowed mask setup in cpuset_init_smp() are removed. The effective_cpus and effective_mems setup there are moved to cpuset_bind(). cc: stable@vger.kernel.org Signed-off-by: Waiman Long <longman@redhat.com> --- kernel/cgroup/cpuset.c | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-)