intel/lkp-tests

Linux Kernel Performance tests

Linux Kernel Performance tests

Getting started

	git clone https://github.com/intel/lkp-tests.git

	cd lkp-tests
	make install

	lkp help

Install dependency packages for jobs

	# browse and select a job you want to run, for example, jobs/hackbench.yaml
	ls lkp-tests/jobs
	
	# install the common dependencies for lkp
	# Note: lkp doesn't support ruby >= 3.0, default ruby version should be 2.x
	lkp install

Run one atomic job

	lkp split-job lkp-tests/jobs/hackbench.yaml
	# output is:
	# lkp-tests/jobs/hackbench.yaml => ./hackbench-pipe-8-process-1600%.yaml
	# lkp-tests/jobs/hackbench.yaml => ./hackbench-socket-8-process-1600%.yaml
	# lkp-tests/jobs/hackbench.yaml => ./hackbench-pipe-8-threads-1600%.yaml
	# lkp-tests/jobs/hackbench.yaml => ./hackbench-socket-8-threads-1600%.yaml
	# lkp-tests/jobs/hackbench.yaml => ./hackbench-pipe-8-process-100%.yaml
	# lkp-tests/jobs/hackbench.yaml => ./hackbench-socket-8-process-100%.yaml
	# lkp-tests/jobs/hackbench.yaml => ./hackbench-pipe-8-threads-100%.yaml
	# lkp-tests/jobs/hackbench.yaml => ./hackbench-socket-8-threads-100%.yaml
	# lkp-tests/jobs/hackbench.yaml => ./hackbench-pipe-8-process-50%.yaml
	# lkp-tests/jobs/hackbench.yaml => ./hackbench-socket-8-process-50%.yaml
	# lkp-tests/jobs/hackbench.yaml => ./hackbench-pipe-8-threads-50%.yaml
	# lkp-tests/jobs/hackbench.yaml => ./hackbench-socket-8-threads-50%.yaml

	# install the remaining dependencies for the splited job
	lkp install ./hackbench-socket-8-threads-50%.yaml
	# or add -f option to force to install all dependencies
	lkp install -f ./hackbench-socket-8-threads-50%.yaml

	lkp run ./hackbench-socket-8-threads-50%.yaml

Run your own disk partitions

Specify disk partitions by defining hdd_partitions/sdd_partitions in host file named with local hostname and then lkp split-job will write the disk partitions information to split job file automatically.

Please note that disk partitions may be formatted/corrupted to run job.

	echo "hdd_partitions: /dev/sda /dev/sdb" >> lkp-tests/hosts/$(hostname | sed -r 's/-[0-9]+$//g' | sed -r 's/-[0-9]+-/-/g')
	lkp split-job lkp-tests/jobs/blogbench.yaml
	# output is:
	# lkp-tests/jobs/blogbench.yaml => ./blogbench-1HDD-ext4.yaml
	# lkp-tests/jobs/blogbench.yaml => ./blogbench-1HDD-xfs.yaml
	# ...
	lkp install ./blogbench-1HDD-ext4.yaml
	lkp run ./blogbench-1HDD-ext4.yaml

Run your own benchmarks

To run your own benchmarks that are not part of lkp-tests, you can use mytest job.

	lkp split-job lkp-tests/jobs/mytest.yaml
	# output is:
	# jobs/mytest.yaml => ./mytest-defaults.yaml
	lkp run ./mytest-defaults.yaml -- <command> <argument> ...

Check result

	lkp result hackbench

Add extra scripts in post run stage

	# create new scripts or rename hidden template scripts in the directory
	echo "echo result_root: \$RESULT_ROOT" > post-run/print-result-root
	lkp run ./ebizzy-10s-1x-200%.yaml
	# output is:
	# ...
	# result_root: /lkp/result/ebizzy/10s-1x-200%/shao2-debian/debian/defconfig/gcc-6/5.7.0-2-amd64/1

Supported Distributions

Most test cases should install/run well in

• Debian sid
• Archlinux
• CentOS7

There is however some initial support for:

• OpenSUSE:
  • jobs/trinity.yaml
• Fedora
• Clear Linux(>=22640)

As for now, lkp-tests still needs to run as root.

Adding distribution support

If you want to add support for your Linux distribution you will need an installer file which allows us to install dependencies per job. For examples look at: distro/installer/* files.

Since packages can have different names we provide an adaptation mapping for a base Ubuntu package (since development started with that) to your own distribution package name, for example adaptation files see: distro/adaptation/*. For now adaptation files must have the architecture dependent packages (ie, that ends with the postfix :i386) towards the end of the adaptation file.

You will also want to add a case for your distribution on sync_distro_sources() on the file lib/install.sh.

Extra Documentation

Refer to https://github.com/intel/lkp-tests/wiki

Contributing

1. Fork it
2. Create your feature branch (git checkout -b my-new-feature)
3. Commit your changes (git commit -am 'Add some feature')
4. Push to the branch (git push origin my-new-feature)
5. Create new Pull Request

lkp-tests's Issues

LKP not cloning & building the benchmark

How to compare two kernel release? (one results copied from another machine)

Hi, I'm benching CentOS 7 and CentOS 8 with lkp-tests on two machines, both have 3 runs. Then copied CentOS 7's result to CentOS 8's:

➜  ~ tree -L 2 /lkp/result/schbench/100%/localhost.localdomain/centos/defconfig/gcc
/lkp/result/schbench/100%/localhost.localdomain/centos/defconfig/gcc
├── 3.10.0-1127.19.1.el7.x86_64
│   ├── 0
│   ├── 1
│   ├── 2
│   ├── avg.csv
│   ├── avg.json
│   ├── desc.yaml
│   ├── matrix.csv
│   ├── matrix.json
│   ├── stddev.csv
│   └── stddev.json
└── 4.18.0-193.19.1.el8_2.x86_64
    ├── 0
    ├── 1
    ├── 2
    ├── avg.csv
    ├── avg.json
    ├── desc.yaml
    ├── matrix.csv
    ├── matrix.json
    ├── stddev.csv
    └── stddev.json

8 directories, 14 files

Then I run:

lkp ncompare -s commit="3.10.0-1127.19.1.el7.x86_64.x86_64" -o commit="4.18.0-193.19.1.el8_2.x86_64" -a

but got nothing.

So I want to know is there a way to compare two release? Thanks!

the fsmark case restart.

 I am a loyal user of LKP, and LKP does provide a lot of convenience for our kernel testing.
 NOW I have a question to ask for advice.
 The scenes is as follows: We have a set of automated distributed execution frameworks from "Patch Crawl, Analysis, Compilation, Test (LKP) to Patch Recommendation", but in the part of Test (LKP),after running the disk cases (fsmark, fio, etc.),I must restart in order to umount the disk,which resulted in sub-cluster lost and can not contact with the scheduler.After the restart,sub-cluster can continue on line, but can not continue to test the case (fsmark) automatic.
 For example: when I split the fsmark case and run the sub-case of it, I must restart in order to unount the disk so that I can continue the next fsmark sub-case testing.  
 Above all, in our system, how to solve the problem of the restart? Whether it can be some settings, can also be tested fsmark case without restart ?

Install lkp-tests dependency on Centos 8 get python34 error

Running lkp install will-it-scale.yaml reported the following error:

...
Use: lkp-tests/distro/installer/centos install openssl-devel procps python3-argcomplete python3-colorama python3-paramiko python34 rsync util-linux
Failed to set locale, defaulting to C.UTF-8
Last metadata expiration check: 0:04:20 ago on Thu Jun 2 14:30:02 2022.
Package openssl-devel-1:1.1.1k-6.el8.x86_64 is already installed.
Package procps-ng-3.3.15-7.el8.x86_64 is already installed.
Package python3-argcomplete-1.9.3-6.el8.noarch is already installed.
No match for argument: python34

Please stop removing changes from master branch history

I just found the change 24a705 disappeared. Please work on feature branches before merging to master branch.
Thank you

Turbostat is x86 only

Going through steps from lkp-tests.pdf I got stopped on 'turbostat':

Making turbostat benchmark for debian                  
timeout 60m git clone -q https://github.com/aaronlu/turbostat.git turbostat
gcc -Wall turbostat.c -c -o /tmp/build-turbostat/turbostat/turbostat.o
turbostat.c:43:10: fatal error: cpuid.h: No such file or directory
 #include <cpuid.h>                                   
          ^~~~~~~~~
compilation terminated.
make: *** [Makefile:14: turbostat.o] Error 1
Making turbostat-LKP failed

Looked at source and noticed that it is x86-64 only tool. And I am on AArch64...

Can turbostat be optional?

can‘t find directory：tools/perf ,while i install the atomic job

     emmm, i am a greenhand, and i completely followed the reference to use lkp. After git clone lkp-test, make install, lkp install, split job,  I installed an atomic job--''./hackbench-50%-threads-socket.yaml'' . I could see from the command line that it showed 'making perf benchmark for centos'. after a while , it failed ,and showed that:

"/root/lkp-tests/pack/perf: line 17: cd: tools/perf: No such file or directory
make: *** No targets specified and no makefile found. Stop.
Making perf-LKP failed"

by the way , i am using tencent CVM , and there is some information here:

top - 22:38:14 up 2 days, 6:38, 1 user, load average: 0.00, 0.01, 0.05
Tasks: 124 total, 1 running, 123 sleeping, 0 stopped, 0 zombie
%Cpu(s): 0.8 us, 0.0 sy, 0.0 ni, 99.2 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
KiB Mem : 16267292 total, 14298988 free, 223200 used, 1745104 buff/cache
KiB Swap: 0 total, 0 free, 0 used. 15689820 avail Mem

I really need your help ,or I will be killed by my boss !!! thanks for your time

rm due to not enough HDD disks

Hi,
I am running iozone benchmark through lkp on X86 (Linux kernel - 5.3.0). I am getting error with HDD disks.

Command: sudo lkp run ./iozone-2HDD-ext4-kyber.yaml
rm due to not enough HDD disks
need: 2, has: 0
Traceback (most recent call last):
6: from /home/test/Ankita/lkp-tests/bin/run-local:131:in

' 5: from /home/test/Ankita/lkp-tests/lib/job.rb:691:in expand_params'
4: from /home/test/Ankita/lkp-tests/lib/job.rb:77:in for_each_in' 3: from /home/test/Ankita/lkp-tests/lib/job.rb:77:in each'
2: from /home/test/Ankita/lkp-tests/lib/job.rb:78:in block in for_each_in' 1: from /home/test/Ankita/lkp-tests/lib/job.rb:694:in block in expand_params'
/home/test/Ankita/lkp-tests/lib/job.rb:669:in `run_filter': /home/test/Ankita/lkp-tests/filters/disk: exitstatus 1 (Job::ParamError)

Please help me with this.
Thanks
Ankita

lkp ncompare using source code

ncompare, _rm broken on latest lkp-tests

One a machine where I have latest build of lkp-tests [head-> 93e6715], I see ncompare being broken,

# lkp ncompare -s commit=5.16.2-v5.16 -s commit=5.15.2-build41 -a  -k
Traceback (most recent call last):
        17: from /home/LKP/lkp-tests/sbin/ncompare:6:in `<main>'
        16: from /home/LKP/lkp-tests/lib/compare.rb:1150:in `compare'
        15: from /home/LKP/lkp-tests/lib/compare.rb:1140:in `parse_argv'
        14: from /home/LKP/lkp-tests/lib/compare.rb:1140:in `map'
        13: from /home/LKP/lkp-tests/lib/compare.rb:1142:in `block in parse_argv'
        12: from /home/LKP/lkp-tests/lib/compare.rb:1142:in `to_a'
        11: from /home/LKP/lkp-tests/lib/nresult_root.rb:391:in `each'
        10: from /home/LKP/lkp-tests/lib/nresult_root.rb:345:in `mrt_table_set'
         9: from /home/LKP/lkp-tests/lib/nresult_root.rb:345:in `new'
         8: from /home/LKP/lkp-tests/lib/nresult_root.rb:278:in `initialize'
         7: from /home/LKP/lkp-tests/lib/nresult_root.rb:248:in `open'
         6: from /home/LKP/lkp-tests/lib/data_store.rb:834:in `open'
         5: from /home/LKP/lkp-tests/lib/data_store.rb:241:in `open'
         4: from /home/LKP/lkp-tests/lib/common.rb:24:in `block in add_cached_method'
         3: from /home/LKP/lkp-tests/lib/common.rb:24:in `new'
         2: from /home/LKP/lkp-tests/lib/data_store.rb:67:in `initialize'
         1: from /home/LKP/lkp-tests/lib/data_store.rb:79:in `load'
/home/LKP/lkp-tests/lib/data_store.rb:84:in `from_data': undefined method `each' for nil:NilClass (NoMethodError)

Same ebizzy results if I remote merge and do a ncompare on another server [where lkp-tests head -> 6ccdc97] it works well:

$ lkp ncompare -s commit=5.16.2-v5.16 -s commit=5.15.2-build41 -a  -k
=========================================================================================
compiler/duration/iterations/kconfig/nr_threads/rootfs/tbox_group/testcase:
  gcc/10s/10x/defconfig/200%/rhel/souparnika/ebizzy

commit:
  5.16.2-v5.16
  5.15.2-build41

5.16.2-v5.16              5.15.2-build41
---------------- ---------------------------
         %stddev     %change         %stddev
             \          |                \
   2884351            -1.2%    2850900        ebizzy.throughput
     18656            -7.3%      17297        ebizzy.throughput.per_thread.max
      7752            +2.9%       7979        ebizzy.throughput.per_thread.min
      0.10           -22.1%       0.08        ebizzy.throughput.per_thread.stddev_percent
  28843510            -1.2%   28509008        ebizzy.workload

I am running on RHEL 8.4, with ruby 2.5.9p229 (2021-04-05 revision 67939) [x86_64-linux]

Add new benchmark to LKP

Hi,
I am trying to add new benchmark to lkp. But after lkp run command getting following error.

Traceback (most recent call last):
5: from /home/ankita/workspace/open_source/lkp-tests/bin/run-local:168:in <main>' 4: from /home/ankita/workspace/open_source/lkp-tests/lib/result_root.rb:380:in convert_mrt'
3: from /home/ankita/workspace/open_source/lkp-tests/lib/result_root.rb:338:in convert_one_mresult_root' 2: from /home/ankita/workspace/open_source/lkp-tests/lib/nresult_root.rb:317:in new_node'
1: from /home/ankita/workspace/open_source/lkp-tests/lib/nresult_root.rb:313:in axes_to_table' /home/ankita/workspace/open_source/lkp-tests/lib/nresult_root.rb:303:in testcase_to_table': Unknow testcase: coremark (RuntimeError)

How do I fix this?

Thanks

Perf compile fails due to missing files

I am porting lkp to Linux raspberrypi4-64 4.19.126-v8 #1 SMP PREEMPT Mon Jun 29 13:37:35 UTC 2020 aarch64 aarch64 aarch64 GNU/Linux.
When I use lkp install ./pft-50%-300s.yaml getting error as
1. mv: cannot stat 'util/.session.o.tmp': No such file or directory
make[4]: *** [/home/root/lkp/lkp-tests/tmp-pkg/perf/src/linux/tools/build/Makefile.build:97: util/session.o] Error 1
make[4]: *** Waiting for unfinished jobs....
make[3]: *** [/home/root/lkp/lkp-tests/tmp-pkg/perf/src/linux/tools/build/Makefile.build:139: util] Error 2
make[2]: *** [Makefile.perf:625: perf-in.o] Error 2
make[1]: *** [Makefile.perf:231: sub-make] Error 2
make: *** [Makefile:70: all] Error 2
==> ERROR: A failure occurred in build().
Aborting...
Install perf-stat failed
2. mv: cannot stat './.builtin-script.o.tmp': No such file or directory
make[3]: *** [/home/root/lkp/lkp-tests/tmp-pkg/perf/src/linux/tools/build/Makefile.build:96: builtin-script.o] Error 1
make[3]: *** Waiting for unfinished jobs....

Please refer attachment for the detailed log.
log.txt

Thanks

Time of run

Hi,
I am running LKP on two different arch, x86 and ARM(RPI4 ) but able to see the difference in ti me for each iteration however it is taking approximately same for x86 but there is variation in time for running LKP on Raspberry pi.
What all factors i should consider to know the exact reason of degradations like why it is taking long ?? Kindly suggest.

Thank you

Errors in job.rb when running job

Hi, the command lkp split jobs/hackbench.yaml gets the following error on my Linux 3.13.0-24-generic x86-64.

/usr/lib/ruby/1.9.1/rubygems/custom_require.rb:36:in `require': /home/luopeng/lkp-tests-master/lib/job.rb:392: unknown type of %string (SyntaxError)
        available_programs %i(workload_elements monitors)
                              ^
/home/luopeng/lkp-tests-master/lib/job.rb:392: syntax error, unexpected $end, expecting keyword_end
        available_programs %i(workload_elements monitors)
                              ^
        from /usr/lib/ruby/1.9.1/rubygems/custom_require.rb:36:in `require'
        from /home/luopeng/lkp-tests-master/sbin/split-job:5:in `<main>'

The lkp install runs well and the dependencies have been installed. So what may be the problem? Thanks.

ERROR: While executing gem ... (NameError) uninitialized constant Gem::RDoc

Hi
I am running lkp on FU540 (linux kernel 5.4.3), when I run lkp split-job jobs/aim9.yaml got following error.
traceback (most recent call last):
11: from /home/root/lkp-tests/sbin/split-job:5:in <main>' 10: from /usr/lib/ruby/2.6.5/rubygems/core_ext/kernel_require.rb:59:in require'
9: from /usr/lib/ruby/2.6.5/rubygems/core_ext/kernel_require.rb:59:in require' 8: from /home/root/lkp-tests/lib/job.rb:6:in <top (required)>'
7: from /usr/lib/ruby/2.6.5/rubygems/core_ext/kernel_require.rb:59:in require' 6: from /usr/lib/ruby/2.6.5/rubygems/core_ext/kernel_require.rb:59:in require'
5: from /home/root/lkp-tests/lib/common.rb:9:in <top (required)>' 4: from /usr/lib/ruby/2.6.5/rubygems/core_ext/kernel_require.rb:59:in require'
3: from /usr/lib/ruby/2.6.5/rubygems/core_ext/kernel_require.rb:59:in require' 2: from /home/root/lkp-tests/lib/array_ext.rb:3:in <top (required)>'
1: from /usr/lib/ruby/2.6.5/rubygems/core_ext/kernel_require.rb:59:in require' /usr/lib/ruby/2.6.5/rubygems/core_ext/kernel_require.rb:59:in require': cannot load such file -- active_support/core_ext/enumerable (LoadError)

I tried to resolve this using gem install activesupport then I got error as,

ERROR: While executing gem ... (NameError)
uninitialized constant Gem::RDoc

Also tried to install rdoc using, gem install rdoc but did not work.

1. How to fix this?
2. Which ruby version lkp require?

Thanks

Alternative of fakeroot

Is there any alternative of fakeroot? Can I use pseudo instead of fakeroot for lkp?

split job error

I got an error in ：lkp split-job lkp-tests/jobs/hackbench.yaml
my env: CentOS release 6.3 (Final)
ruby:2.3 how to resolve this problem,can you help me @MaXinjian

/root/ruby/lib/ruby/2.3.0/rubygems/core_ext/kernel_require.rb:55:in require': cannot load such file -- active_support/core_ext/enumerable (LoadError) from /root/ruby/lib/ruby/2.3.0/rubygems/core_ext/kernel_require.rb:55:in require'
from /home/bae/lkp/lkp-tests/lib/array_ext.rb:3:in <top (required)>' from /root/ruby/lib/ruby/2.3.0/rubygems/core_ext/kernel_require.rb:55:in require'
from /root/ruby/lib/ruby/2.3.0/rubygems/core_ext/kernel_require.rb:55:in require' from /home/bae/lkp/lkp-tests/lib/common.rb:9:in <top (required)>'
from /root/ruby/lib/ruby/2.3.0/rubygems/core_ext/kernel_require.rb:55:in require' from /root/ruby/lib/ruby/2.3.0/rubygems/core_ext/kernel_require.rb:55:in require'
from /home/bae/lkp/lkp-tests/lib/job.rb:6:in <top (required)>' from /root/ruby/lib/ruby/2.3.0/rubygems/core_ext/kernel_require.rb:55:in require'
from /root/ruby/lib/ruby/2.3.0/rubygems/core_ext/kernel_require.rb:55:in require' from /home/bae/lkp/lkp-tests/sbin/split-job:5:in

Add ipc-bench benchmark to lkp

Hi,
I am trying to add ipc-bench benchmark to lkp.
I tried to add ipc-bench.yaml for the benchmark in lkp-tests/jobs/ directory. When I used lkp split-job jobs/ipc-bench.yaml I get split job as ipc-bench-defaults.yaml , not with parameters that I have added in ipc-bench.yaml file. How do I fix this?
Could you please help me to understand 'how to add new benchmark to lkp-tests'.

Thanks

Feature: Support comparing the performance against 2 different devices

Please consider to add a feature for comparing the performance against 2 different devices. Such as, comparing the KPIs from 2 block devices with lkp ncompare.
Thanks!

Originally posted by @SCHEN2015 in https://github.com/intel/lkp-tests/issue_comments#issuecomment-425688598

Run job.yaml failed in Xsever

I installed lkp-tests on Ubuntu 18.04 machine, and run jobs.yaml on it to reproduce "https://patchwork.kernel.org/patch/11780373/" patch regression issue.
But I got the following Xserver error, could anyone help to check how to resolve it on jobs.yaml. Thanks,

~/lkp-tests$ sudo lkp run job.yaml
2020-10-07 17:17:26
for cpu_dir in /sys/devices/system/cpu/cpu[0-9]*
do
online_file="$cpu_dir"/online
[ -f "$online_file" ] && [ "$(cat "$online_file")" -eq 0 ] && continue

file="$cpu_dir"/cpufreq/scaling_governor
[ -f "$file" ] && echo "performance" > "$file"

done

IPMI BMC is not supported on this machine, skip bmc-watchdog setup!

_XSERVTransSocketUNIXCreateListener: ...SocketCreateListener() failed
_XSERVTransMakeAllCOTSServerListeners: server already running
(EE)
Fatal server error:
(EE) Cannot establish any listening sockets - Make sure an X server isn't already running(EE)
(EE)
Please consult the The X.Org Foundation support
at http://wiki.x.org
for help.
(EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
(EE)
(EE) Server terminated with error (1). Closing log file.
XIO: fatal IO error 11 (Resource temporarily unavailable) on X server ":0"
after 7 requests (7 known processed) with 0 events remaining.

ltp benchmark parameters

Hi

I am running ltp benchmark through lkp (HiFive Unleashed (RISCV) yocto linux kernel 5.4.3 with 8GB RAM). Only 6GB is available when running LKP. I find that lkp is expecting the ram size of the 8 GB to run the ltp.
Here is the error:

rm due to not enough memory
need: 8192 MB, has: 6866 MB

And with HDD disk jobs
This is the error:

rm due to not enough HDD disks
need: 1, has: 0

Can we get more information on the test execution as we need to validate LTP through LKP framework on RISCV?

Regards
Megha

isn't a atomic jobfile, only atomic jobfile is supported

Hello
I am running lkp for Linux raspberrypi4-64 5.4.51-v8 #1 SMP PREEMPT Mon Jul 20 05:13:46 UTC 2020 aarch64 aarch64 aarch64 GNU/Linux. After lkp split-job & lkp install, when I run a job using lkp run ./aim9-add_double-300s.yaml get following error,
2020-07-21 08:51:27 +0000 ERROR -- Looks ./aim9-add_double-300s.yaml isn't a atomic jobfile, only atomic jobfile is supported
2020-07-21 08:51:27 +0000 ERROR -- Please run 'lkp split-job ./aim9-add_double-300s.yaml' first

How do I fix this?

Thanks

ERROR: One or more files did not pass the validity check! Install aim9 failed

Hi,
I am getting Install aim9 failed error with aim9-lkp run.

aim9::
==> Making package: aim9 1-1 (Mon Dec 28 10:28:48 IST 2020)
==> Checking runtime dependencies...
==> Checking buildtime dependencies...
==> Retrieving sources...
-> Downloading s9110.tar.Z...
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 397 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0
100 375 100 375 0 0 207 0 0:00:01 0:00:01 --:--:-- 0
100 138 100 138 0 0 60 0 0:00:02 0:00:02 --:--:-- 60
100 310 100 310 0 0 87 0 0:00:03 0:00:03 --:--:-- 522
100 168k 100 168k 0 0 20738 0 0:00:08 0:00:08 --:--:-- 48623
==> WARNING: Skipping verification of source file PGP signatures.
==> Validating source files with md5sums...
s9110.tar.Z ... FAILED
==> ERROR: One or more files did not pass the validity check!
Install aim9 failed

How do I fix this?

Thanks

Failure when using output created by "lkp compile lkp-test/tests/phoronix-test-suite.yaml"

Steps to reproduce

1. phoronix-test-suite try to connect to the Internet, if use http proxy, please run 'phoronix-test-suite network-setup'

2. Create an auto execution bash script

    $ lkp compile lkp-tests/tests/phoronix-test-suite.yaml > phoronix-test-suite.sh
    $ chmod a+x phoronix-test-suite.sh
    $ mkdri -p result
    $ yes 2 | sudo lkp run -o result/ phoronix-test-suite.sh

2. The execution will only run the first test suite, in this case it is "aio-stress-1.1.1"

3. Refer to the attach for the quick hack fix phoronix-test-suite.zip

4. This type issue exists in most of the tests that use the output by using "lkp compile ..." command. It also has problem with input parameters that not numeric, e.g. 50%, 12.5% and 100%.

Question about using `lkp ncompare`

I'm so sorry to disturb you. But after studying many articles (blog/pdf/slides/README.md) I still can not get lkp ncompare to run successfully.

For example: in https://01.org/zh/node/15555 it says:
# lkp ncompare -s commit=9e66317d3c92ddaab330c125dfe9d06eee268aff -o commit=a11c3148ba6b8b4cac4876b7269d570260d0bd41
But I have no idea about where to get the commit id.

Does lkp ncompare command support comparing other things beside the kernel? for example, compare 2 versions of a driver or compare 2 different NVMe devices?
Such as: After testing fio-basic against 2 devices, is there any available tools to tell the performance difference?

[root@ip-10-22-2-226 lkp]# lkp result fio-basic
/lkp/result/fio-basic/10s-8t-1SSD-xfs-read-4k-sync-5g/ip-10-22-2-226.us-west-2.compute.internal/rhel/x86_64-rhel-7.2/gcc-7/3.10.0-862.el7.x86_64/0/
/lkp/result/fio-basic/10s-8t-1SSD-xfs-read-4k-sync-5g/ip-10-22-2-226.us-west-2.compute.internal/rhel/x86_64-rhel-7.2/gcc-7/3.10.0-862.el7.x86_64/1/

Thank you in advance!

Split Job appears broken

With the latest master branch, trying to reproduce a test fails on the split-job test:

bin/lkp split-job --compatible job.yaml
Traceback (most recent call last):
	3: from kp-tests/sbin/split-job:95:in `<main>'
	2: from lkp-tests/sbin/split-job:95:in `each'
	1: from kp-tests/sbin/split-job:98:in `block in <main>'
lkp-tests/lib/job.rb:202:in `load': wrong number of arguments (given 2, expected 1) (ArgumentError)

This seems to have been introduced in 9e624cd as checking out before that commit seems to work fine.

I'm running on a Debian Bullseye system with ruby 2.7.4p191

Result of chromeswap benchmark

Hi,
Can I get sample output of chromeswap benchmark? I need it for reference.

Thanks

Cannot find "install" command under /{bin,sbin,tools,lkp-exec}/ on raspberrypi

Hi,
I am facing following error on raspberrypi.

/lkp-tests# lkp install
readlink: invalid option -- 'e'
BusyBox v1.34.1 () multi-call binary.

Usage: readlink [-fnv] FILE
BusyBox v1.34.1 () multi-call binary.

Usage: dirname FILENAME
BusyBox v1.34.1 () multi-call binary.

Usage: dirname FILENAME
Cannot find "install" command under /{bin,sbin,tools,lkp-exec}/

What could be the reason?

Thanks

ebizzy test run error

Following to documentation passage to run the ebizzy test (lkp install, split-job, job install and job run; everything with sudo) cause the following output:

<path-to>/lkp-tests/monitors/turbostat: 12: cd: can't cd to /lkp/benchmarks/turbostat
<path-to>/lkp-tests/tests/ebizzy: 15: cd: can't cd to /lkp/benchmarks/ebizzy

If I create the two directory, then I get:

Iteration: 1
PWD: /lkp/benchmarks/ebizzy
2021-03-09 21:20:33 ./ebizzy -t 4 -S 10
/home/andrea/Documents/test-definitions/automated/linux/lkp/lkp-tests/tests/ebizzy: 21: ./ebizzy: not found
[repeated for 100 iteration]

I modified the script at /home/andrea/Documents/test-definitions/automated/linux/lkp/lkp-tests/tests/ebizzy to print the current directory, and as you can see, it's /lkp/benchmarks/ebizzy

I am on latest version of Ubuntu. I run hackbench test and it seems to run correctly.

Feature: Merge results from different machines

#20 (comment)

Please consider to add a feature for lkp ncompare to deal with the results from different machines. Thanks!

NVML:Compilation error with clang compiler

root@SSP-CLXAEP-CDI205/home/labuser/drill/lkp-tests # lkp run ./nvml-arch-none.yaml
2019-06-24 04:40:25 cd /lkp/benchmarks/nvml/src/test
2019-06-24 04:40:25 export CC=clang
2019-06-24 04:40:25 export CXX=clang++
2019-06-24 04:40:25 cd /lkp/benchmarks/nvml
2019-06-24 04:40:25 export CC=clang
2019-06-24 04:40:25 export CXX=clang++
2019-06-24 04:40:25 cp src/test/testconfig.sh.example src/test/testconfig.sh
2019-06-24 04:40:25 make EXTRA_CFLAGS=-DPAGE_SIZE=4096 -DUSE_VALGRIND USE_LLVM_LIBCPP=1 LIBCPP_INCDIR=/usr/local/libcxx/include/c++/v1/ LIBCPP_LIBDIR=/usr/local/libcxx/lib
make -C src all
make[1]: Entering directory '/lkp/benchmarks/nvml/src'
make -C libpmem
make[2]: Entering directory '/lkp/benchmarks/nvml/src/libpmem'
clang-8: error: unknown argument: '-ftree-loop-distribute-patterns'
clang-8: error: unknown argument: '-fno-semantic-interposition'
clang-8: warning: -Wl,-z: 'linker' input unused [-Wunused-command-line-argument]
clang-8: warning: -Wl,now: 'linker' input unused [-Wunused-command-line-argument]
clang-8: warning: -Wl,-z: 'linker' input unused [-Wunused-command-line-argument]
clang-8: warning: -Wl,relro: 'linker' input unused [-Wunused-command-line-argument]
clang-8: warning: -Wl,-sort-common: 'linker' input unused [-Wunused-command-line-argument]
clang-8: warning: -Wl,--enable-new-dtags: 'linker' input unused [-Wunused-command-line-argument]
clang-8: warning: optimization flag '-ffat-lto-objects' is not supported [-Wignored-optimization-argument]
../Makefile.inc:178: *** "clang -g -O3 -feliminate-unused-debug-types -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=32 -Wformat -Wformat-security -m64 -fasynchronous-unwind-tables -Wp,-D_REENTRANT -ftree-loop-distribute-patterns -Wl,-z -Wl,now -Wl,-z -Wl,relro -fno-semantic-interposition -ffat-lto-objects -fno-signed-zeros -fno-trapping-math -fassociative-math -Wl,-sort-common -Wl,--enable-new-dtags -mtune=skylake -std=gnu99 -fno-common -pthread -DSRCVERSION="1.6+git330.g57a94c062" -DSDS_ENABLED -DNDCTL_GE_63 -DPAGE_SIZE=4096 -DUSE_VALGRIND -dM -E -x c /dev/null -o /dev/null" failed. Stop.
make[2]: Leaving directory '/lkp/benchmarks/nvml/src/libpmem'
make[1]: *** [Makefile:162: libpmem] Error 2
make[1]: Leaving directory '/lkp/benchmarks/nvml/src'
make: *** [Makefile:82: all] Error 2
build env failed
/home/labuser/drill/lkp-tests/lib/debug.sh:12: die
/home/labuser/drill/lkp-tests/tests/nvml:15: main
:0:
root@SSP-CLXAEP-CDI205/home/labuser/drill/lkp-tests # uname -a
Linux SSP-CLXAEP-CDI205 5.1.12-785.native #1 SMP Wed Jun 19 07:09:10 UTC 2019 x86_64 GNU/Linux.
Workaround:-
By commenting out export CC=clang,export CXX=clang++( In the path /lib/tests/nvml.sh ,in setup_compiler function),able to build successfully without errors.

Testing blocked by an issue from time.rb

While testing with CentOS 7.5, I encountered the following issue which has blocked the report from generating.

Description

/usr/share/ruby/time.rb:264:in `local': argument out of range (ArgumentError)

Debug Log

......
wait for background monitors: 22739 22791 22793 22854 22799 22804 22895 22839 22810 22889 uptime proc-vmstat proc-stat diskstats meminfo slabinfo sched_debug softirqs interrupts turbostat
/usr/share/ruby/time.rb:264:in `local': argument out of range (ArgumentError)
	from /usr/share/ruby/time.rb:264:in `make_time'
	from /usr/share/ruby/time.rb:331:in `parse'
	from /root/lkp/lkp-tests/stats/mpstat:25:in `parse_time'
	from /root/lkp/lkp-tests/stats/mpstat:44:in `block in <main>'
	from /root/lkp/lkp-tests/stats/mpstat:37:in `each_line'
	from /root/lkp/lkp-tests/stats/mpstat:37:in `<main>'
2018-09-18 07:41:27 ERROR -- /root/lkp/lkp-tests/stats/mpstat /lkp/result/hackbench/50%-threads-socket/ip-10-22-2-38.us-west-2.compute.internal/centos/x86_64-rhel-7.2/gcc-7/4.9.112-32.el7.x86_64/0/mpstat < /lkp/result/hackbench/50%-threads-socket/ip-10-22-2-38.us-west-2.compute.internal/centos/x86_64-rhel-7.2/gcc-7/4.9.112-32.el7.x86_64/0/mpstat  exit code 1, check /dev/shm/lkp-stats.RGn0CrLi

And there is no outputs can be found in /dev/shm/lkp-stats.RGn0CrLi.

How reproducible

100%

Step to reproduce

lkp install --ssd /dev/xvdb --hdd /dev/xvdc
lkp split-job lkp-tests/jobs/hackbench.yaml
lkp install ./hackbench-50%-threads-socket.yaml
lkp run ./hackbench-50%-threads-socket.yaml

Version-Release number of selected component

[root@ip-10-22-2-38 lkp]# cat /etc/system-release
CentOS Linux release 7.5.1804 (Core) 
[root@ip-10-22-2-38 lkp]# uname -r
4.9.112-32.el7.x86_64
[root@ip-10-22-2-38 lkp]# ruby --version
ruby 2.0.0p648 (2015-12-16) [x86_64-linux]
[root@ip-10-22-2-38 lkp]# gem list

*** LOCAL GEMS ***

ast (2.4.0)
bigdecimal (1.2.0)
builder (3.2.3)
bundler (1.16.4)
ci_reporter (2.0.0)
ci_reporter_rspec (1.0.0)
coderay (1.1.2)
colorize (0.8.1)
diff-lcs (1.3)
docile (1.3.1)
git (1.5.0)
gnuplot (2.6.2)
io-console (0.4.2)
json (2.1.0, 1.7.7)
method_source (0.9.0)
parallel (1.12.1)
parser (2.5.1.2)
powerpack (0.1.2)
pry (0.11.3)
pry-doc (0.13.4)
psych (2.0.0)
rainbow (2.2.2)
rake (12.3.1)
rdoc (4.0.0)
rspec (3.8.0)
rspec-core (3.8.0)
rspec-expectations (3.8.1)
rspec-mocks (3.8.0)
rspec-support (3.8.0)
rubocop (0.50.0)
ruby-progressbar (1.10.0)
simplecov (0.16.1)
simplecov-html (0.10.2)
simplecov-rcov (0.2.3)
term-ansicolor (1.6.0)
tins (1.16.3)
unicode-display_width (1.4.0)
yard (0.9.16)

Additional information

1. This problem can also be reproduced on RHEL7.5.

PACKERDRILL:Compilation Error

root@SSP-CLXAEP-CDI205/home/labuser/drill/lkp-tests # lkp install ./packetdrill-defaults.yaml
Use: /home/labuser/drill/lkp-tests/distro/installer/clear install bc os-core os-core-update sysadmin-basic c-basic dev-utils dev-utils-dev git linux-tools network-basic openssl os-core-update os-testsuite-0day patch perl-extras ruby-basic
Successfully installed bundler-2.0.2
Parsing documentation for bundler-2.0.2
Done installing documentation for bundler after 1 seconds
1 gem installed
Don't run Bundler as root. Bundler can ask for sudo if it is needed, and installing your bundle as root will break this application for all non-root
users on this machine.
Using rake 12.3.3
Using concurrent-ruby 1.1.5
...........................
.
.
.............
gcc -g -Wall -Werror -c -o run_packet.o run_packet.c
run_packet.c: In function ‘find_tcp_timestamp’:
run_packet.c:408:13: error: taking address of packed member of ‘struct tcp_option’ may result in an unaligned pointer value [-Werror=address-of-packed-member]
408 | (void *)&(option->data.time_stamp.val);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
run_packet.c:410:13: error: taking address of packed member of ‘struct tcp_option’ may result in an unaligned pointer value [-Werror=address-of-packed-member]
410 | (void *)&(option->data.time_stamp.ecr);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
cc1: all warnings being treated as errors
make: *** [: run_packet.o] Error 1
Making packetdrill-LKP failed
root@SSP-CLXAEP-CDI205/home/labuser/drill/lkp-tests # uname -a
Linux SSP-CLXAEP-CDI205 5.1.12-785.native #1 SMP Wed Jun 19 07:09:10 UTC 2019 x86_64 GNU/Linux
Workaround :-
It seems like there is an incompatibility between packetdrill test suite and gcc9 compiler – compiler complains about an alignment of a pointer in packed struct. the alignment of the pointer value should not be related to the alignment of the pointer variable itself.
For now, commenting out the werror flag (In the path ::/tmp/build-packetdrill/packetdrill/gtests/net/packetdrill/Makefile.common) and able to build packetdrill test packages successfully without errors.

Can I use this in ARM64 for Android?

Hi
From the Read me seems ARM64 is not said support, is there a possible to run it on ARM64 android phone ?
BRs,
Chao

+0530 WARN -- Jobfile is a dirty file, better to run 'lkp split-job --compatible ./chromeswap-5-uniform-0.25.yaml' first

Hi
could you check this?

$ lkp run ./chromeswap-5-uniform-0.25.yaml
2021-08-09 16:18:34 +0530 WARN -- Jobfile is a dirty file, better to run 'lkp split-job --compatible ./chromeswap-5-uniform-0.25.yaml' first

bsdtar is not installable in recent Ubuntu releases

The bsdtar binary package is installed, using apt, when running lkp install on Ubuntu. This fails since Ubuntu 19.10 as the bsdtar binary package no longer exists. The bsdtar binary can be found in the libarchive-tools binary package in most, but not all, currently supported Ubuntu releases. For lkp install to work across all currently supported Ubuntu releases, we need bsdtar to be installed on some Ubuntu releases and libarchive-tools to be installed on others. Is there a way to specify that in the distro adaptation files?

Here's the breakdown of bsdtar to libarchive-tools across all currently supported Ubuntu releases:

Ubuntu Release	bsdtar	libarchive-tools
Ubuntu Focal	no	yes
Ubuntu 19.10 (Eoan)	no	yes
Ubuntu 19.04 (Disco)	yes[1]	yes
Ubuntu 18.04 LTS (Bionic)	yes[1]	yes
Ubuntu 16.04 LTS (Xenial)	yes	no

[1] bsdtar exists as a transitional binary package that depends on libarchive-tools so installing bsdtar will cause libarchive-tools to also be installed

I think the same problem exists in Debian but I haven't tested this. According to the bsdtar page on packages.debian.org, it looks to me as if bsdtar cannot be installed on Debian unstable so the next bullseye release will not have an installable bsdtar binary package.

NUC device shut down while running "phoronix-test-suite run supertuxkart-1.5.2"

I installed Ubuntu 18.04.4 LTS (kernel: v5.4) on Intel NUC8, and installed lkp-tests.

But when I run "phoronix-test-suite run supertuxkart-1.5.2" on NUC8, then system always goes shut down soon. Therefore, I can't fully complete the performance test to validate my upstream patch (https://patchwork.kernel.org/project/intel-gfx/patch/[email protected]/)

Please help on this urgent issue. Thanks,

The log is as attached.
run_log.txt

Cannot install some packages in lkp-tests/distro/depends/lkp

Hi,
I am trying to port lkp-tests for customized distro & and I need that updated sources should work on both x86 & customized distro.
The updated sources are working fine with customized distro but after running lkp install on ubuntu(X86) get following error,
$ sudo lkp install
Cannot install some packages in lkp-tests/distro/depends/lkp

How do I fix this?

Thanks

Unable to execute autotest-part1-cyclictest on Clear Linux

Hello,

I am a new user of lkp-tests and have been playing with example tests. Well the ones that I used worked on Ubuntu 18.04, but when shifted to Clear Linux. The lkp-tests for the out of the box example autotest-part1-cyclictest.yaml failed. Here are the steps executed:

lkp split jobs/autotest-part1.yaml - PASSED
sudo lkp install autotest-part1-cyclictest.yaml - PASSED
sudo lkp run autotest-part1-cyclictest.yaml - Failed with the following error message:
2019-07-25 22:10:47 python2.7 client/autotest-local run cyclictest
22:10:48 ERROR| Can not find test cyclictest/control
kill 18106 vmstat --timestamp -n 10
kill 18104 dmesg --follow --decode
wait for background processes: 18109 meminfo

The same test works fine in Ubuntu with kernel version: 4.15.0-54-generic, but fails on Clear Linux with kernel version: 4.19.50-12.iot-lts2018-preempt-rt

lkp ncompare result interpetation

Hi,

I am new to lkp test suite.I have installed it and found useful in comparing different kernel versions.

Can somebody help me to understand it better way.Attached is the lkp compare test result.Column headings are unclear to me.

What does 5322 and 379 mean in the attached result?How can I say which kernel is good from the result?

lkp ncompare -s commit="3.10.0-1062.9.1.el7.x86_64" -o commit="3.18.0-rc5" -a
======================= ==================================================================
compiler/ipc/kconfig/mode/nr_threads/rootfs/tbox_group/testcase:
gcc-7/pipe/x86_64-rhel-7.6/process/1600%/centos/bn1dhcp244-49.mvista.com/hackbench

commit:
3.10.0-1062.9.1.el7.x86_64
3.18.0-rc5

3.10.0-1062.9.1. 3.18.0-rc5

---

   fail:runs  %reproduction    fail:runs
       |             |             |    
   5322:1      -494373%         379:1     dmesg.timestamp:last
   5322:1      -494373%         379:1     kmsg.timestamp:last

Thanks in Advance!!!

unixbench: libperl5.24 does not exist on ubuntu-18.04

ubuntu-18.04 repos seem to only contain libperl5.26 and perl-modules-5.26, but unixbench attempts to install 5.24. Is ubuntu 18.04 not a supported distro? If it is supported, it seems that we will need a way to specify to install whichever version of the package is available. Ideally, it would be nice to be able to specify something like "perl-modules >= 5.26", but the current method of installing packages doesn't seem to be equipped to handle that sort of thing.

perf: Error: unknown switch `c'

lkp may need changes to adapt to perf from Linux 5.1:

root@33c9bf576e93:~/lkp-tests# lkp run  ./hackbench-socket-threads-50%.yaml
2019-03-28 10:46:37 +0000 WARN -- skip non-executable /root/lkp-tests/stats/build-ceph
/root/lkp-tests/monitors/vmstat: 5: /root/lkp-tests/monitors/vmstat: vmstat: not found
/root/lkp-tests/monitors/heartbeat: 5: /root/lkp-tests/monitors/heartbeat: vmstat: not found
/root/lkp-tests/monitors/heartbeat: 23: exec: vmstat: not found
/root/lkp-tests/monitors/vmstat: 23: exec: vmstat: not found
/root/lkp-tests/monitors/mpstat: 27: /root/lkp-tests/monitors/mpstat: /lkp/benchmarks/sysstat/usr/local/bin/mpstat: not found                                                                  
dmesg: read kernel buffer failed: Operation not permitted
dmesg: read kernel buffer failed: Operation not permitted
/root/lkp-tests/monitors/perf-stat: 89: /root/lkp-tests/monitors/perf-stat: cannot create /proc/sys/kernel/nmi_watchdog: Read-only file system                                                 
  Error: unknown switch `c'

 Usage: perf stat [<options>] [<command>]

root@33c9bf576e93:~/lkp-tests# /lkp/benchmarks/perf/perf --version
perf version 5.1.rc2.g14c741

PIGLIT:Compilation error

root@hyd-clx-coffeelake/home/testrunner/lkp_packet/lkp-tests # lkp install ./piglit-all-igt-all.yaml
Use: /home/testrunner/lkp_packet/lkp-tests/distro/installer/clear install bc os-core os-core-update sysadmin-basic c-basic dev-utils dev-utils-dev git linux-tools network-basic openssl os-core-update os-testsuite-0day patch perl-extras ruby-basic
Error: Curl - Communicating with server timed out
Error: Failed to connect to update server: https://cdn.download.clearlinux.org/update
Error: Failed updater initialization. Exiting now
Loading required manifests...
Warning: Bundle "ruby-basic" is already installed, skipping it...
1 bundle was already installed
Successfully installed bundler-2.0.2
Parsing documentation for bundler-2.0.2
Done installing documentation for bundler after 1 seconds
1 gem installed
Don't run Bundler as root. Bundler can ask for sudo if it is needed, and installing your bundle as root will break this application for all non-root
users on this machine.
Using rake 12.3.3
Using concurrent-ruby 1.1.5
Using i18n 1.6.0
..............
..................
.............
make -C /home/testrunner/lkp_packet/lkp-tests/bin/event
Use: /home/testrunner/lkp_packet/lkp-tests/distro/installer/clear install c-basic desktop-apps openssl os-core-dev os-core-update os-testsuite-0day sysadmin-basic
Making makepkg benchmark for clear
makepkg is already installed into /lkp/benchmarks/makepkg, skip this packing
Installing makepkg-LKP
Just make /lkp/benchmarks/makepkg-x86_64.cgz
Use: /home/testrunner/lkp_packet/lkp-tests/distro/installer/clear install desktop dev-utils-dev libva-utils os-core-dev python-extras qt-basic-dev xfce4-desktop c-basic c-extras-gcc7 cpp-7 dev-utils dev-utils-dev libgcc1 libgcc-7-dev libmpfr6 libva-utils os-clr-on-clr-dev os-core os-core-dev python2-basic python-extras qt-basic-dev
Loading required manifests...
Warning: Bundle "cpp-7" is invalid, skipping it...
Failed to install 1 of 1 bundles
Loading required manifests...
Warning: Bundle "libgcc1" is invalid, skipping it...
Failed to install 1 of 1 bundles
Loading required manifests...
Warning: Bundle "libgcc-7-dev" is invalid, skipping it...
Failed to install 1 of 1 bundles
Loading required manifests...
Warning: Bundle "libmpfr6" is invalid, skipping it...
Failed to install 1 of 1 bundles
Making piglit benchmark for clear
timeout 60m git remote update origin
Fetching origin
timeout 60m git checkout -q origin/master
timeout 60m git remote update origin
Fetching origin
timeout 60m git checkout -q origin/master
timeout 60m git remote update origin
Fetching origin
timeout 60m git checkout -q origin/master
/home/testrunner/lkp_packet/lkp-tests/pack/piglit: line 27: [: : integer expression expected
[ 1%] Built target cmocka
[ 2%] Built target threads_bundled
[ 3%] Building C object src/waffle/CMakeFiles/waffle-1.dir/wayland/wayland_display.c.o
In file included from /tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:32:
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_wrapper.h:99:32: error: conflicting types for ‘wfl_wl_display_roundtrip’
99 | #define wl_display_roundtrip (*wfl_wl_display_roundtrip)
| ^~~~~~~~~~~~~~~~~~~~~~~~
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_wrapper.h:61:3: note: previous declaration of ‘wfl_wl_display_roundtrip’ was here
61 | (*wfl_wl_display_roundtrip)(struct wl_display *display);
| ^~~~~~~~~~~~~~~~~~~~~~~~
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c: In function ‘registry_listener_global’:
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:72:31: error: implicit declaration of function ‘wl_registry_bind’ [-Werror=implicit-function-declaration]
72 | self->wl_compositor = wl_registry_bind(self->wl_registry, name,
| ^~~~~~~~~~~~~~~~
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:72:29: error: assignment to ‘struct wl_compositor *’ from ‘int’ makes pointer from integer without a cast [-Werror=int-conversion]
72 | self->wl_compositor = wl_registry_bind(self->wl_registry, name,
| ^
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:76:24: error: assignment to ‘struct wl_shell *’ from ‘int’ makes pointer from integer without a cast [-Werror=int-conversion]
76 | self->wl_shell = wl_registry_bind(self->wl_registry, name,
| ^
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c: At top level:
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:88:21: error: variable ‘registry_listener’ has initializer but incomplete type
88 | static const struct wl_registry_listener registry_listener = {
| ^~~~~~~~~~~~~~~~~~~~
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:89:6: error: ‘const struct wl_registry_listener’ has no member named ‘global’
89 | .global = registry_listener_global,
| ^~~~~~
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:89:15: warning: excess elements in struct initializer
89 | .global = registry_listener_global,
| ^~~~~~~~~~~~~~~~~~~~~~~~
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:89:15: note: (near initialization for ‘registry_listener’)
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:90:6: error: ‘const struct wl_registry_listener’ has no member named ‘global_remove’
90 | .global_remove = registry_listener_global_remove
| ^~~~~~~~~~~~~
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:90:22: warning: excess elements in struct initializer
90 | .global_remove = registry_listener_global_remove
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:90:22: note: (near initialization for ‘registry_listener’)
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c: In function ‘wayland_display_connect’:
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:111:25: error: implicit declaration of function ‘wl_display_get_registry’; did you mean ‘wl_display_get_global’? [-Werror=implicit-function-declaration]
111 | self->wl_registry = wl_display_get_registry(self->wl_display);
| ^~~~~~~~~~~~~~~~~~~~~~~
| wl_display_get_global
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:111:23: error: assignment to ‘struct wl_registry *’ from ‘int’ makes pointer from integer without a cast [-Werror=int-conversion]
111 | self->wl_registry = wl_display_get_registry(self->wl_display);
| ^
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:117:13: error: implicit declaration of function ‘wl_registry_add_listener’; did you mean ‘wl_keyboard_add_listener’? [-Werror=implicit-function-declaration]
117 | error = wl_registry_add_listener(self->wl_registry,
| ^~~~~~~~~~~~~~~~~~~~~~~~
| wl_keyboard_add_listener
In file included from /tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:32:
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c: In function ‘wayland_display_sync’:
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_wrapper.h:99:31: error: void value not ignored as it ought to be
99 | #define wl_display_roundtrip (*wfl_wl_display_roundtrip)
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:184:9: note: in expansion of macro ‘wl_display_roundtrip’
184 | if (wl_display_roundtrip(dpy->wl_display) == -1) {
| ^~~~~~~~~~~~~~~~~~~~
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c: At top level:
/tmp/build-piglit/waffle/src/waffle/wayland/wayland_display.c:88:42: error: storage size of ‘registry_listener’ isn’t known
88 | static const struct wl_registry_listener registry_listener = {
| ^~~~~~~~~~~~~~~~~
cc1: some warnings being treated as errors
make[2]: *** [src/waffle/CMakeFiles/waffle-1.dir/build.make:492: src/waffle/CMakeFiles/waffle-1.dir/wayland/wayland_display.c.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:592: src/waffle/CMakeFiles/waffle-1.dir/all] Error 2
make: *** [Makefile:152: all] Error 2
Making piglit-LKP failed
root@hyd-clx-coffeelake/home/testrunner/lkp_packet/lkp-tests # uname -a
Linux hyd-clx-coffeelake 4.19.1-656.native #1 SMP Tue Nov 13 16:11:01 UTC 2018 x86_64 GNU/Linux

Bug: libhugetlbfs-test multiple pagenums

https://github.com/intel/lkp-tests/blob/master/jobs/libhugetlbfs-test.yaml
lists both 32 and 64 for pagenums.
When that is fed into this (at least on rpm distros):
https://github.com/intel/lkp-tests/blob/master/tests/libhugetlbfs-test#L14
it results in the command including --add-temp-swap=32 64 --pool-pages-min which is not proper syntax to hugeadm. Should the test loop on pagenum value or something instead?

nr_threads becomes 0, rm

Hi,
I am getting following error, what could be the reason?
$ lkp run ./hackbench-pipe-10-process-50%.yaml
nr_threads becomes 0, rm
Traceback (most recent call last):
6: from /home/root/lkp/lkp-tests/bin/run-local:131:in <main>' 5: from /home/root/lkp/lkp-tests/lib/job.rb:691:in expand_params'
4: from /home/root/lkp/lkp-tests/lib/job.rb:77:in for_each_in' 3: from /home/root/lkp/lkp-tests/lib/job.rb:77:in each'
2: from /home/root/lkp/lkp-tests/lib/job.rb:78:in block in for_each_in' 1: from /home/root/lkp/lkp-tests/lib/job.rb:694:in block in expand_params'
/home/root/lkp/lkp-tests/lib/job.rb:669:in `run_filter': /home/root/lkp/lkp-tests/filters/nr_threads: exitstatus 2 (Job::ParamError)

Thanks

Unable to add arugments to a new job file with cyclictest

The examples for cyclictest under jobs/autotest-part1 do not have examples with parameters passed. While out-of-box tests for cyclictest when executed from autotest-part1(autotest-part1-cyclictest.yaml) work fine.

Example cyclictest without arugments which works fine is autotest-cyclictest.yaml:
suite: autotest
testcase: autotest
category: functional

autotest:
test:

• cyclictest

I see in the test results that the cyclictest ran with args: -t 10 -l 100000
I want to run cyclictest with parameters: -S -p 99 -d 0 -m -D 10s -h 400 -q

What are the steps to follow to be able to execute a test with the above parameters.

+0530 WARN -- Jobfile is a dirty file, better to run 'lkp split-job --compatible ./vm-scalability-300s-small-allocs.yaml' first

Hi,
How to resolve below error?

$ sudo lkp run ./vm-scalability-300s-small-allocs.yaml
2021-07-15 18:56:05 +0530 WARN -- Jobfile is a dirty file, better to run 'lkp split-job --compatible ./vm-scalability-300s-small-allocs.yaml' first

$ sudo lkp split-job --compatible ./vm-scalability-300s-small-allocs.yaml
/tmp/20210715-20631-szgfe9 => ./20210715-20631-szgfe9-300s-small-allocs.yaml

$ sudo lkp run ./20210715-20631-szgfe9-300s-small-allocs.yaml
2021-07-15 18:56:37 +0530 WARN -- Jobfile is a dirty file, better to run 'lkp split-job --compatible ./20210715-20631-szgfe9-300s-small-allocs.yaml' first

Thanks

unixbench: install issue on ubuntu-20.04 and ubuntu-21.04

ubuntu21.04 and ubuntu 20.04 repos seem to contain perl-modules-5.32 and perl-modules-5.30 instead of perl-modules-5.26 which is what unixbench dependencies try to install via lkp install. Similarly ubuntu21.04 and ubuntu 20.04 repos seem to contain libperl5.32 and libperl5.30 instead of libperl5.26 which is currently present in distro/adaptation/ubuntu.

I tried removing version information from these two packages distro/adaptation/ubuntu,thinking it would pick latest available version of the packages, but that does not seem to work incase of libperl. Since same adaptation file is used in case of both ubuntu versions, if I change the version for one release another release seem to fail.

No package libunwind8 available in Fedora 28/25

[root@7cd76314c885 lkp-tests]# cat /etc/redhat-release 
Fedora release 28 (Twenty Eight)

Also verified this issue in Fedora 25.

diff --git a/distro/adaptation/fedora b/distro/adaptation/fedora
index a0bbdbc1..2e5d1231 100644
--- a/distro/adaptation/fedora
+++ b/distro/adaptation/fedora
@@ -128,3 +128,4 @@ xz-utils: xz
 zlib1g-dev: zlib-devel
 protobuf-compiler: protobuf
 libprotobuf-dev: protobuf-devel
+libunwind8: libunwind

May I contribute upon code into intel/lkp-tests?

An issue about ftrace has blocked the testing on CentOS 7.5

While testing with CentOS 7.5, I encountered the following issue which has blocked the report from generating.

Description:
2018-09-26 08:19:34 ERROR -- /root/lkp/lkp-tests/sbin/dump-stat ftrace exit code 1, check /lkp/result/dd-write/10m-1HDD-cfq-xfs-1dd/ip-10-22-2-38.us-west-2.compute.internal/centos/x86_64-rhel-7.2/gcc-7/4.9.112-32.el7.x86_64/0/ftrace.*.yaml

[root@ip-10-22-2-38 lkp]# lkp run dd-write-1hdd-10m-1HDD-cfq-xfs-1dd.yaml
......
(standard_in) 2: syntax error
/root/lkp/lkp-tests/monitors/ftrace: line 37: echo: write error: Invalid argument
/root/lkp/lkp-tests/monitors/ftrace: line 37: echo: write error: Invalid argument
/root/lkp/lkp-tests/monitors/ftrace: line 37: echo: write error: Invalid argument
/root/lkp/lkp-tests/monitors/ftrace: line 37: echo: write error: Invalid argument
cp: omitting directory ‘/tmp/lkp-root/trace_event’
/root/lkp/lkp-tests/monitors/ftrace: line 45: echo: write error: Invalid argument
2018-09-26 08:09:27 timeout -s KILL 600  dd  if=/dev/zero of=/fs/xvdc/zero-1 status=noxfer &
......
/root/lkp/lkp-tests/stats/wrapper: line 161: /lkp/result/dd-write/10m-1HDD-cfq-xfs-1dd/ip-10-22-2-38.us-west-2.compute.internal/centos/x86_64-rhel-7.2/gcc-7/4.9.112-32.el7.x86_64/0/ftrace.*.yaml: No such file or directory
2018-09-26 08:19:34 ERROR -- /root/lkp/lkp-tests/sbin/dump-stat ftrace  exit code 1, check /lkp/result/dd-write/10m-1HDD-cfq-xfs-1dd/ip-10-22-2-38.us-west-2.compute.internal/centos/x86_64-rhel-7.2/gcc-7/4.9.112-32.el7.x86_64/0/ftrace.*.yaml

Step to reproduce:

lkp install --ssd /dev/xvdb --hdd /dev/xvdc
lkp split-job lkp-tests/jobs/dd-write-1hdd.yaml
lkp install ./dd-write-1hdd-10m-1HDD-cfq-xfs-1dd.yaml
lkp run ./dd-write-1hdd-10m-1HDD-cfq-xfs-1dd.yaml

How reproducible:
100%

Additional information:

1. This problem can also be reproduced on RHEL7.5.
2. Ruby and Gem versions:

[root@ip-10-22-2-38 lkp]# ruby --version
ruby 2.0.0p648 (2015-12-16) [x86_64-linux]
[root@ip-10-22-2-38 lkp]# gem list
*** LOCAL GEMS ***
ast (2.4.0)
bigdecimal (1.2.0)
builder (3.2.3)
bundler (1.16.4)
ci_reporter (2.0.0)
ci_reporter_rspec (1.0.0)
coderay (1.1.2)
colorize (0.8.1)
diff-lcs (1.3)
docile (1.3.1)
git (1.5.0)
gnuplot (2.6.2)
io-console (0.4.2)
json (2.1.0, 1.7.7)
method_source (0.9.0)
parallel (1.12.1)
parser (2.5.1.2)
powerpack (0.1.2)
pry (0.11.3)
pry-doc (0.13.4)
psych (2.0.0)
rainbow (2.2.2)
rake (12.3.1)
rdoc (4.0.0)
rspec (3.8.0)
rspec-core (3.8.0)
rspec-expectations (3.8.1)
rspec-mocks (3.8.0)
rspec-support (3.8.0)
rubocop (0.50.0)
ruby-progressbar (1.10.0)
simplecov (0.16.1)
simplecov-html (0.10.2)
simplecov-rcov (0.2.3)
term-ansicolor (1.6.0)
tins (1.16.3)
unicode-display_width (1.4.0)
yard (0.9.16)