build-configs-cip.yaml: Disable arm64 builds for 4.4-cip

[fbezdeka]

arm64 was out of scope for the 4.4 series. Disabling it save some
resources and silences arm64 related testing issues.

Signed-off-by: Florian Bezdeka florian.bezdeka@siemens.com

[fbezdeka]

@aliceinwire, according to last CIP meeting I'm trying to disable arm64 builds for 4.4-cip only. Let's see what the test run on staging reports...

[patersonc]

I'm guessing we need to trigger a build manually on staging for the cip 4.4 kernel?

Latest staging build was a couple of weeks ago: https://staging.kernelci.org/job/cip/

[aliceinwire]

started cip_4.4 staging test
The results will appear here: https://staging.kernelci.org/job/cip/

[fbezdeka]

Thanks for starting the build.

Seems we have 5 findings now:

• baseline/x86_64_defconfig+x86-chromebook
  https://storage.staging.kernelci.org/cip/linux-4.4.y-cip/v4.4.302-cip70-97-g3635ffa3f2b2/x86_64/x86_64_defconfig+x86-chromebook/gcc-10/lab-collabora/baseline-hp-x360-14-G1-sona.html
  looks like the prompt was not properly detected
• baseline-cip-nfs/beaglebone-black multi_v7_defconfig
  https://storage.staging.kernelci.org/cip/linux-4.4.y-cip/v4.4.302-cip70-97-g3635ffa3f2b2/arm/multi_v7_defconfig/gcc-10/lab-broonie/baseline-cip-nfs-beaglebone-black.html
  No idea what went wrong here. Seems we reached the auto-login and tried to fire some shell commands. Unclear why a u-boot action is marked as failed here
• kselftest-alsa/hp-x360-14-G1-sona x86_64_defconfig+x86-chromebook
  Seems like a failure in trying to get NFS rootfs mounted, no interfaces available? Missing driver?
  https://storage.staging.kernelci.org/cip/linux-4.4.y-cip/v4.4.302-cip70-97-g3635ffa3f2b2/x86_64/x86_64_defconfig+x86-chromebook/gcc-10/lab-collabora/kselftest-alsa-hp-x360-14-G1-sona.html
• ltp-fcntl-locktests/asus-C523NA-A20057-coral x86_64_defconfig+x86-chromebook
  https://storage.staging.kernelci.org/cip/linux-4.4.y-cip/v4.4.302-cip70-97-g3635ffa3f2b2/x86_64/x86_64_defconfig+x86-chromebook/gcc-10/lab-collabora/ltp-fcntl-locktests-asus-C523NA-A20057-coral.html
  Looks identical to the previous one. Missing driver?
• ltp-pty/asus-C523NA-A20057-coral x86_64_defconfig+x86-chromebook
  https://storage.staging.kernelci.org/cip/linux-4.4.y-cip/v4.4.302-cip70-97-g3635ffa3f2b2/x86_64/x86_64_defconfig+x86-chromebook/gcc-10/lab-collabora/ltp-pty-asus-C523NA-A20057-coral.html
  Looks identical to the previous one. Missing driver?

[aliceinwire]

looks like arm64 is correctly disabled.
@fbezdeka please don't mix pull request off topic discussion here.
open a new issue or pr for that problems.

[aliceinwire]

@gctucker can you check this?

[gctucker]

It appears to be working as expected according to the results on the web dashboard. I'm actually slightly surprised as it's the first time we use filters this way, normally build configurations are defined "positively" with a set of configs to build for a particular branch rather than "negatively" with a full config with a filter to remove some.

But I've double-checked that locally using this command:

$ ./kci_build list_kernel_configs --build-config=cip_4.4 --kdir=/home/gtucker/src/linux | sort > configs-4.4.txt

First with the actual linux-4.4.y branch checked out in the linux kernel tree, then the exact same command but with linux-4.9 (which doesn't normally happen in production but that's a detail). Here's the interesting part of the diff:

$ diff -u configs-4.4.txt configs-4.9.txt 
--- configs-4.4.txt	2022-08-18 19:26:54.040003265 +0200
+++ configs-4.9.txt	2022-08-18 19:30:49.648003732 +0200
@@ -9,16 +9,22 @@
 arc tinyconfig gcc-10
 arc vdk_hs38_defconfig gcc-10
 arc vdk_hs38_smp_defconfig gcc-10
+arc zebu_hs_defconfig gcc-10
+arc zebu_hs_smp_defconfig gcc-10
+arm64 allnoconfig gcc-10
+arm64 defconfig gcc-10
+arm64 tinyconfig gcc-10
 arm acs5k_defconfig gcc-10
 arm acs5k_tiny_defconfig gcc-10
 arm allnoconfig gcc-10
 arm am200epdkit_defconfig gcc-10

The arm64 configs are present with 4.9 but not with 4.4. So the filter is working as expected. Nice.