==== Meeting 2023/12/13 ==== === Attendees === Jan-Simon, Walt, Scott, Marius ===Topics=== * lava lab status * 11/29 * Lab-baylibre operational * Lab-agl-core down due to system update * LAVA update available, need to schedule uprev with Ryan * 12/13 * Maint window scheduled with Ryan * uprevs/ -next branch / Denys * 11/29 * 3.1.29 uploaded * 4.0.14 is WIP * -next is waiting for the next cycle * 12/13 * Next look after CES * Yocto-check-layer - https://jira.automotivelinux.org/browse/SPEC-4680 * 10/18 * no update * 11/29 * Next step is then to deal with meta-agl-demo and all its dependencies. This is a larger operation and unsure we can actually deal with all dependency layers. We need to see if we can exempt external layers and only check meta-agl-demo content in the meantime. * Extend Testing * 11/29 * No update * IC image and integration of drm-leasemanager * 11/2 * Scott has image locally for drm-lease, * lxc network is broken for agl-cluster-demo-lxc-host.bb * focus on agl-host-drm-lease-test image * 11/29 * Images merged https://gerrit.automotivelinux.org/gerrit/c/AGL/meta-agl-devel/+/29355 * fixme's: order of displays is not fixed, network issues * need to work out automated test * CI to build image * LAVA test description, shell script * New Beagle Boards * Denys pushed an update. * No update * 12/13 * Templates for BeagleAI-64 and BeaglePlay * meta-ti update in gerrit * Pi 5 * No update (e.g meta-raspberrypi) * icu-native change in https://gerrit.automotivelinux.org/gerrit/c/AGL/meta-agl-demo/+/29486 * symptom: it does not trigger a rebuild in https://build.automotivelinux.org/job/ci-platform-meta-agl-demo-verify-CIB-html5-qemux86-64/409/consoleFull as it should. (grep for icu-native) * https://jira.automotivelinux.org/browse/SPEC-4980 * https://bugzilla.yoctoproject.org/show_bug.cgi?id=15302 Next meeting: Jan 17th ==== Meeting 2023/11/29 ==== === Attendees === Jan-Simon, Walt, Scott, Yan, Marius ===Topics=== * lava lab status * 11/29 * Lab-baylibre operational * Lab-agl-core down due to system update * LAVA update available, need to schedule uprev with Ryan * uprevs/ -next branch / Denys * 11/2 * e.g. after workshop we can take a look again. * 11/29 * 3.1.29 uploaded * 4.0.14 is WIP * -next is waiting for the next cycle * Yocto-check-layer - https://jira.automotivelinux.org/browse/SPEC-4680 * 10/18 * no update * 11/29 * Next step is then to deal with meta-agl-demo and all its dependencies. This is a larger operation and unsure we can actually deal with all dependency layers. We need to see if we can exempt external layers and only check meta-agl-demo content in the meantime. * Extend Testing * 11/29 * No update * * IC image and integration of drm-leasemanager * 11/2 * Scott has image locally for drm-lease, * lxc network is broken for agl-cluster-demo-lxc-host.bb * focus on agl-host-drm-lease-test image * 11/29 * Images merged https://gerrit.automotivelinux.org/gerrit/c/AGL/meta-agl-devel/+/29355 * fixme's: order of displays is not fixed, network issues * need to work out automated test * CI to build image * LAVA test description, shell script * RISC-V * rust/ring crate situation getting better, databroker supposed to build now * new release of databroker available gerrit 29490 * BeagleAI * Denys pushed an update. * No update * Pi 5 * No update (e.g meta-raspberrypi) New: * icu-native change in https://gerrit.automotivelinux.org/gerrit/c/AGL/meta-agl-demo/+/29486 * symptom: it does not trigger a rebuild in https://build.automotivelinux.org/job/ci-platform-meta-agl-demo-verify-CIB-html5-qemux86-64/409/consoleFull as it should. (grep for icu-native) * bitbake -S printdiff icu-native * Need debug session w/ Ryan Next meeting: Dec 13th ==== Meeting 2023/11/02 ==== === Attendees === Jan-Simon, Walt, Scott ===Topics=== * lava lab status * 11/02 * Lab Baylibre operational * AGL core lab up * Ryan scheduled Updates and announced. * jenkins not executing jobs due to an migration issue of the cloud configuration * jenkins is back up now * bbe build failing. need to debug. * * -next branch / Denys * 10/18 * no news, waiting for update cycle after Nanbield release. * Next LTS will be YP 5.0 Scarthgap . Release is planned for Apr. * 11/2 * e.g. after workshop we can take a look again. * Yocto-check-layer - https://jira.automotivelinux.org/browse/SPEC-4680 * 10/18 * no update * Extend Testing * 10/18 * * IC image and integration of drm-leasemanager * 09/20 * Next: bring up simple demo image with 2 containers bringing up a static reference image * 10/04 * Scott looking into the container images next * 10/18 * WIP * 11/2 * Scott has image locally for drm-lease, * lxc network is broken for agl-cluster-demo-lxc-host.bb * focus on agl-host-drm-lease-test image * RISC-V * rust/ring crate situation getting better, databroker supposed to build now * BeagleAI * Denys pushed an update. * Pi 5 * first boards shipping, sourced 2 * am335x (bbe) gfx works again ... New: Next meeting: Nov 15th ==== Meeting 2023/10/04 ==== === Attendees === Jan-Simon, Yan, Marius, Scott, Denys ===Topics=== * lava lab status * 10/18 * Lab Baylibre operational * AGL core lab up * Ryan scheduled Updates and announced. * -next branch / Denys * 10/04 * mesa issue fixed, but still visible in -next branch (depends on kernel) * good news is all is building atm, * waiting for next update cycle from upstream YP * 10/18 * no news, waiting for update cycle after Nanbield release. * Next LTS will be YP 5.0 Scarthgap . Release is planned for Apr. * Yocto-check-layer - https://jira.automotivelinux.org/browse/SPEC-4680 * 10/04 * no update * 10/18 * no update * Extend Testing * 10/04 * 10/18 * * IC image and integration of drm-leasemanager * 09/20 * Next: bring up simple demo image with 2 containers bringing up a static reference image * 10/04 * Scott looking into the container images next * 10/18 * WIP New: * discussed -next, RiscV and rust/crates <-- revisit * RiscV + rust (ring crate) to build * discussed BeaglePlay or BeagleAI ? Will look at BeagleAI first. * will have about 4 weeks to find solution for BeagleAI * Pi5 will get 3-4 boards starting 10/23 for workshop * * am335x: https://git.yoctoproject.org/meta-ti/commit/?id=c3b08caa9b63ea287963c9d02a448e00b3bb7653 * gfx could work again Next meeting: Nov 1st ==== Meeting 2023/10/04 ==== === Attendees === Jan-Simon, Walt, Scott, Denys, Marius ===Topics=== * lava lab status * 10/04 RiscV board being moved into production lab@baylibre, config upstreamed * Lab Baylibre operational * AGL core lab * Likely need a series of system uprev's across infra. * -next branch / Denys * 10/04 * mesa issue fixed, but still visible in -next branch (depends on kernel) * good news is all is building atm, * waiting for next update cycle from upstream YP * Yocto-check-layer - https://jira.automotivelinux.org/browse/SPEC-4680 * 10/04 * no update * Extend Testing * 10/04 * IC image and integration of drm-leasemanager * 09/20 * Next: bring up simple demo image with 2 containers bringing up a static reference image * 10/04 * Scott looking into the container images next New: * discussed -next, RiscV and rust/crates <-- revisit * RiscV + rust (ring crate) to build * discussed BeaglePlay or BeagleAI ? Will look at BeagleAI first. * will have about 4 weeks to find solution for BeagleAI * Pi5 will get 3-4 boards starting 10/23 for workshop * Next meeting: Oct 18th ==== Meeting 2023/09/20 ==== === Attendees === Jan-Simon, Walt, Scott, Yan, ===Topics=== * lava lab status * 09/06 One RiscV board arrived @ baylibre * 09/20 Upload of initial board config * Lab Baylibre operational * AGL core lab down due to software update and issues with lava-docker * -next branch / Denys * 09/06 * Denys: try a full agl-ivi-demo-platform build on -next so we can verify the next YP release does work. * 09/20 * Denys is working on the next branch builds. * Yocto-check-layer - https://jira.automotivelinux.org/browse/SPEC-4680 * 08/09 * Expansion on hold till after release of pike. * Extend Testing * 09/20 * IC image and integration of drm-leasemanager * 05/31 * Summary: patched qemu version hosted on docker, can bring up screens. * Next: bring up simple demo image with 2 containers bringing up a static reference image * 06/14 * no update * 08/09 * Next: bring up simple demo image with 2 containers bringing up a static reference image * Artifacts for IC demo -> 3b) - for releases for pi4 * 09/6 * WIP * 09/20 * Next: bring up simple demo image with 2 containers bringing up a static reference image * Artifacts for IC demo -> 3b) - for releases for pi4 * Working on segfault during build: https://build.automotivelinux.org/job/release-pike-icdemo-snapshot/MACHINE=qemux86-64,label=agl-test-slave/2/console New: * discussed -next, RiscV and rust/crates * discussed BeaglePlay or BeagleAI ? Will look at BeagleAI first. Next meeting: Oct 4th ==== Meeting 2023/09/06 ==== === Attendees === Jan-Simon, Walt, Scott, Marius ===Topics=== * lava lab status * 09/06 One RiscV board en-route * -next branch / Denys * 08/09 * Uninative bug (patchelf) affected master (4.0.11 uprev). * test with new uninative 4.1 is here https://build.automotivelinux.org/job/test-single-uninative/AGLBRANCH=master,MACHINE=qemux86-64,label=agl-test-slave/5/console * Richard reported a warning: https://autobuilder.yoctoproject.org/typhoon/#/builders/120/builds/3139 * 09/06 * Denys: try a full agl-ivi-demo-platform build on -next so we can verify the next YP release does work. * Yocto-check-layer - https://jira.automotivelinux.org/browse/SPEC-4680 * 08/09 * Expansion on hold till after release of pike. * 09/06 * Extend Testing * IC image and integration of drm-leasemanager * 05/31 * Summary: patched qemu version hosted on docker, can bring up screens. * Next: bring up simple demo image with 2 containers bringing up a static reference image * 06/14 * no update * 08/09 * Next: bring up simple demo image with 2 containers bringing up a static reference image * Artifacts for IC demo -> 3b) - for releases for pi4 * 09/06 New: Next meeting: Sept 20th ==== Meeting 2023/08/09 ==== === Attendees === Jan-Simon, Scott, Marius ===Topics=== * lava lab status * uprev done (server, core-lab, baylibre) * got 2 sifive unleashed, setup WIP * -next branch / Denys * 06/14 * Updates WIP, close to done * Keep a closer eye starting in september timeframe on -next (2-3 - weekly) so we can fix issues in the upcoming release. * 8/9 * Uninative bug (patchelf) affected master (4.0.11 uprev). * test with new uninative 4.1 is here https://build.automotivelinux.org/job/test-single-uninative/AGLBRANCH=master,MACHINE=qemux86-64,label=agl-test-slave/5/console * Richard reported a warning: https://autobuilder.yoctoproject.org/typhoon/#/builders/120/builds/3139 * Yocto-check-layer - https://jira.automotivelinux.org/browse/SPEC-4680 * 8/9 * Expansion on hold till after release of pike. * Extend Testing * Upload fixed by corentin * https://lava.automotivelinux.org/scheduler/job/33397 * http://135.181.99.107:8002/artifacts/home/agl/2023/05/30/15/36/ * https://gerrit.automotivelinux.org/gerrit/c/src/qa-testdefinitions/+/28935 * recommend NO_COLOR=1 for pytest runs to avoid color escape sequences in logs. * Discussed https://gerrit.automotivelinux.org/gerrit/c/AGL/meta-agl-devel/+/28858 * IC image and integration of drm-leasemanager * rcar + kvm + pi4 * qemux86-64 * rcar config set to one output on each drm lease * AI1: qemu-x86-64 commandline how to boot the ic-multicontainer image (Damian) * Marius max_outputs=2 == two outputs in qemu * AI2: port above configuration to lava and enable image boot (Jan-Simon) * AI3: extend test to take snapshots within container guests (Jan-Simon and Damian) * AI4: add rcar3 build (kingfisher or refhw tbd) * 02/28 * working on qemu cmdline for multiple/two outputs * 03/22 * Problem: we can setup multiple ports, but they are not connected. Patch sent upstream, but no movement so far. * AI: confirm with Corentin if we can use a dockerized 'custom' qemu * Marius offers to reply with a tested-by: * 04/05 * AI: confirm with Corentin if we can use a dockerized 'custom' qemu * YES: https://lava.baylibre.com/scheduler/job/3473045/definition#defline97 * Option 1: use 1-2 vnc-clients to connect to target and bring screen up - problem needs to connect before cluster starts to connect the output. * Option 2: Custom qemu in docker would already run as patch enables 2nd output by default. * AI(jsmoeller): create qemu with patch and host on docker to test * https://gitlab.com/qemu-project/qemu/-/issues/1107 * WIP version of kernelci/qemu with our patches * AI(Damian): qemu cmdline example (Option2 first , Option1 good for dev) * -device virtio-vga,max_outputs=2,id=vid -display vnc=unix:,display=vid,head=0 -display vnc=unix:,display=vid,head=1 * plus simple client tool to open * 05/03: * Can bring up qemu with 2 screens, problem is connecting from the outside still. * Custom qemu-docker WIP. bullseye has qemu-5.2 , would need patch for that version. * https://lava.automotivelinux.org/scheduler/job/33185 * job with correct option turned on https://lava.automotivelinux.org/scheduler/job/33187 * connect_outputs=on needs to be set * working! * AI: stripped-down homescreen to display reference image only, alternative is weston-image in fullscreen mode * reference image: https://git.automotivelinux.org/src/qa-testdefinitions/tree/reference-images * 05/31 * Summary: patched qemu version hosted on docker, can bring up screens. * Next: bring up simple demo image with 2 containers bringing up a static reference image * 06/14 * no update * 08/09 * Next: bring up simple demo image with 2 containers bringing up a static reference image * Artifacts for IC demo -> 3b) - for releases for pi4 * SPEC-4783 -> dnsmasq in KVM demo images conflict with netboot * inspect flutter, html5 and ic-multicontainer builds wrt sstate reuse - done. Times down. * flutter still a bit high * html5 looks good, flutter still high, qemux86-64 a bit longer than expected. New: * build host uprev pending to 20.04 or 22.04 * docs for pike Next meeting: Sept 6 ==== Meeting 2023/06/14 ==== === Attendees === Jan-Simon, Walt, Yan, Scott, Marius ===Topics=== * lava lab status * uprev done (server, core-lab, baylibre) * got 2 sifive unleashed, need to setup * -next branch / Denys * 05/31 * lets bump next before M1 * 06/14 * Updates WIP, close to done * Keep a closer eye starting in september timeframe on -next (2-3 - weekly) so we can fix issues in the upcoming release. * Yocto-check-layer - https://jira.automotivelinux.org/browse/SPEC-4680 * 04/19 * meta-agl-devel layers left * 06/14 * jenkins will run a YCL run for the layers we got working for all CI builds * meta-agl-devel layers left * Extend Testing * Upload fixed by corentin * https://lava.automotivelinux.org/scheduler/job/33397 * http://135.181.99.107:8002/artifacts/home/agl/2023/05/30/15/36/ * https://gerrit.automotivelinux.org/gerrit/c/src/qa-testdefinitions/+/28935 * recommend NO_COLOR=1 for pytest runs to avoid color escape sequences in logs. * Discussed https://gerrit.automotivelinux.org/gerrit/c/AGL/meta-agl-devel/+/28858 * IC image and integration of drm-leasemanager * rcar + kvm + pi4 * qemux86-64 * rcar config set to one output on each drm lease * AI1: qemu-x86-64 commandline how to boot the ic-multicontainer image (Damian) * Marius max_outputs=2 == two outputs in qemu * AI2: port above configuration to lava and enable image boot (Jan-Simon) * AI3: extend test to take snapshots within container guests (Jan-Simon and Damian) * AI4: add rcar3 build (kingfisher or refhw tbd) * 02/28 * working on qemu cmdline for multiple/two outputs * 03/22 * Problem: we can setup multiple ports, but they are not connected. Patch sent upstream, but no movement so far. * AI: confirm with Corentin if we can use a dockerized 'custom' qemu * Marius offers to reply with a tested-by: * 04/05 * AI: confirm with Corentin if we can use a dockerized 'custom' qemu * YES: https://lava.baylibre.com/scheduler/job/3473045/definition#defline97 * Option 1: use 1-2 vnc-clients to connect to target and bring screen up - problem needs to connect before cluster starts to connect the output. * Option 2: Custom qemu in docker would already run as patch enables 2nd output by default. * AI(jsmoeller): create qemu with patch and host on docker to test * https://gitlab.com/qemu-project/qemu/-/issues/1107 * WIP version of kernelci/qemu with our patches * AI(Damian): qemu cmdline example (Option2 first , Option1 good for dev) * -device virtio-vga,max_outputs=2,id=vid -display vnc=unix:,display=vid,head=0 -display vnc=unix:,display=vid,head=1 * plus simple client tool to open * 05/03: * Can bring up qemu with 2 screens, problem is connecting from the outside still. * Custom qemu-docker WIP. bullseye has qemu-5.2 , would need patch for that version. * https://lava.automotivelinux.org/scheduler/job/33185 * job with correct option turned on https://lava.automotivelinux.org/scheduler/job/33187 * connect_outputs=on needs to be set * working! * AI: stripped-down homescreen to display reference image only, alternative is weston-image in fullscreen mode * reference image: https://git.automotivelinux.org/src/qa-testdefinitions/tree/reference-images * 05/31 * Summary: patched qemu version hosted on docker, can bring up screens. * Next: bring up simple demo image with 2 containers bringing up a static reference image * 06/14 * no update * Rahul (AMD): benchmarking the OS, gfxbench, 3dmark, * Marius: glmark2 (wayland-es backend works) * https://layers.openembedded.org/layerindex/recipe/5190/ * gfxbench or 3dmark are more featureful tests * there are no recipes * xwayland required * Unigine has downloadable installer, but only x86. Commandline-driven version is not free. * LD_TRACE_LOADED_OBJECTS=1 /usr/bin/foo to find missing libraries * Between Scott, Marius & Jan-Simon check if xwayland can be enabled * Note: talk with upstream wrt enabling xwayland w/o x11 DISTRO feature * 05/03 * could build xwayland (enabled x11 distro-feature), but had to workaround qtwayland issue. * need change in compositor to start xwayland during startup * Mahdi: let's talk about a container build env for devel and ci * SPEC-4783 -> dnsmasq in KVM demo images conflict with netboot * inspect flutter, html5 and ic-multicontainer builds wrt sstate reuse - done. Times down. * flutter still a bit high * packagegroup rework - https://jira.automotivelinux.org/browse/SPEC-4813 * merged in master New: - Error: Transaction test error: file /etc/polkit-1/rules.d conflicts between attempted installs of polkit-rule-agl-app-1.0-r0.82.aarch64 and polkit-0.119-r0.83.aarch64 -- could we use a different path to install the rule ? -- pseudo bug ?? -- JIRA - Flutter image -> see buildstat wrt re-use and spdx here https://build.automotivelinux.org/job/ci-platform-AGL-repo-verify-CIB-flutter-qemux86-64/70/consoleFull Next meeting: July 26th ! ==== Meeting 2023/05/31 ==== === Attendees === Jan-Simon, Walt, Yan, Scott, Damian, Ishii-san, Marius, Kurokawa-san ===Topics=== * lava lab status * uprev done (server, core-lab, baylibre) * h3 firmware update * -next branch / Denys * 05/03 * planning a bump of next layer * 05/31 * lets bump next before M1 * Yocto-check-layer - https://jira.automotivelinux.org/browse/SPEC-4680 * 03/22 * patches to meta-agl-demo merged * 04/19 * meta-agl-devel layers left * Extend Testing * 04/19 * Artifactorial being restarted, including pyartiproxy * Upload fixed by corentin * https://lava.automotivelinux.org/scheduler/job/33397 * http://135.181.99.107:8002/artifacts/home/agl/2023/05/30/15/36/ * https://gerrit.automotivelinux.org/gerrit/c/src/qa-testdefinitions/+/28935 * recommend NO_COLOR=1 for pytest runs to avoid color escape sequences in logs. * IC image and integration of drm-leasemanager * rcar + kvm + pi4 * qemux86-64 * rcar config set to one output on each drm lease * AI1: qemu-x86-64 commandline how to boot the ic-multicontainer image (Damian) * Marius max_outputs=2 == two outputs in qemu * AI2: port above configuration to lava and enable image boot (Jan-Simon) * AI3: extend test to take snapshots within container guests (Jan-Simon and Damian) * AI4: add rcar3 build (kingfisher or refhw tbd) * 02/28 * working on qemu cmdline for multiple/two outputs * 03/22 * Problem: we can setup multiple ports, but they are not connected. Patch sent upstream, but no movement so far. * AI: confirm with Corentin if we can use a dockerized 'custom' qemu * Marius offers to reply with a tested-by: * 04/05 * AI: confirm with Corentin if we can use a dockerized 'custom' qemu * YES: https://lava.baylibre.com/scheduler/job/3473045/definition#defline97 * Option 1: use 1-2 vnc-clients to connect to target and bring screen up - problem needs to connect before cluster starts to connect the output. * Option 2: Custom qemu in docker would already run as patch enables 2nd output by default. * AI(jsmoeller): create qemu with patch and host on docker to test * https://gitlab.com/qemu-project/qemu/-/issues/1107 * WIP version of kernelci/qemu with our patches * AI(Damian): qemu cmdline example (Option2 first , Option1 good for dev) * -device virtio-vga,max_outputs=2,id=vid -display vnc=unix:,display=vid,head=0 -display vnc=unix:,display=vid,head=1 * plus simple client tool to open * 05/03: * Can bring up qemu with 2 screens, problem is connecting from the outside still. * Custom qemu-docker WIP. bullseye has qemu-5.2 , would need patch for that version. * https://lava.automotivelinux.org/scheduler/job/33185 * job with correct option turned on https://lava.automotivelinux.org/scheduler/job/33187 * connect_outputs=on needs to be set * working! * AI: stripped-down homescreen to display reference image only, alternative is weston-image in fullscreen mode * reference image: https://git.automotivelinux.org/src/qa-testdefinitions/tree/reference-images * 05/31 * Summary: patched qemu version hosted on docker, can bring up screens. * Next: bring up simple demo image with 2 containers bringing up a static reference image * Rahul (AMD): benchmarking the OS, gfxbench, 3dmark, * Marius: glmark2 (wayland-es backend works) * https://layers.openembedded.org/layerindex/recipe/5190/ * gfxbench or 3dmark are more featureful tests * there are no recipes * xwayland required * Unigine has downloadable installer, but only x86. Commandline-driven version is not free. * LD_TRACE_LOADED_OBJECTS=1 /usr/bin/foo to find missing libraries * Between Scott, Marius & Jan-Simon check if xwayland can be enabled * Note: talk with upstream wrt enabling xwayland w/o x11 DISTRO feature * 05/03 * could build xwayland (enabled x11 distro-feature), but had to workaround qtwayland issue. * need change in compositor to start xwayland during startup * Mahdi: let's talk about a container build env for devel and ci * SPEC-4783 -> dnsmasq in KVM demo images conflict with netboot new: * inspect flutter, html5 and ic-multicontainer builds wrt sstate reuse * packagegroup rework - https://jira.automotivelinux.org/browse/SPEC-4813 Next meeting: June 14th ==== Meeting 2023/05/03 ==== === Attendees === Jan-Simon, Walt, Scott, Damian, Marius ===Topics=== * lava lab status * uprev done (server, core-lab, baylibre) * lava server container did not restart * fixed by reinstalling backup * operational but some logs need to be restored still * -next branch / Denys * 04/05 * no update * 04/19 * no update * 05/03 * planning a bump of next layer * Yocto-check-layer - https://jira.automotivelinux.org/browse/SPEC-4680 * 03/22 * patches to meta-agl-demo merged * 04/19 * meta-agl-devel layers left * Extend Testing * 04/19 * Artifactorial being restarted, including pyartiproxy * * IC image and integration of drm-leasemanager * rcar + kvm + pi4 * qemux86-64 * rcar config set to one output on each drm lease * AI1: qemu-x86-64 commandline how to boot the ic-multicontainer image (Damian) * Marius max_outputs=2 == two outputs in qemu * AI2: port above configuration to lava and enable image boot (Jan-Simon) * AI3: extend test to take snapshots within container guests (Jan-Simon and Damian) * AI4: add rcar3 build (kingfisher or refhw tbd) * 02/28 * working on qemu cmdline for multiple/two outputs * 03/22 * Problem: we can setup multiple ports, but they are not connected. Patch sent upstream, but no movement so far. * AI: confirm with Corentin if we can use a dockerized 'custom' qemu * Marius offers to reply with a tested-by: * 04/05 * AI: confirm with Corentin if we can use a dockerized 'custom' qemu * YES: https://lava.baylibre.com/scheduler/job/3473045/definition#defline97 * Option 1: use 1-2 vnc-clients to connect to target and bring screen up - problem needs to connect before cluster starts to connect the output. * Option 2: Custom qemu in docker would already run as patch enables 2nd output by default. * AI(jsmoeller): create qemu with patch and host on docker to test * https://gitlab.com/qemu-project/qemu/-/issues/1107 * WIP version of kernelci/qemu with our patches * AI(Damian): qemu cmdline example (Option2 first , Option1 good for dev) * -device virtio-vga,max_outputs=2,id=vid -display vnc=unix:,display=vid,head=0 -display vnc=unix:,display=vid,head=1 * plus simple client tool to open * 05/03: * Can bring up qemu with 2 screens, problem is connecting from the outside still. * Custom qemu-docker WIP. bullseye has qemu-5.2 , would need patch for that version. * https://lava.automotivelinux.org/scheduler/job/33185 * job with correct option turned on https://lava.automotivelinux.org/scheduler/job/33187 * connect_outputs=on needs to be set * working! * AI: stripped-down homescreen to display reference image only, alternative is weston-image in fullscreen mode * reference image: https://git.automotivelinux.org/src/qa-testdefinitions/tree/reference-images * Rahul (AMD): benchmarking the OS, gfxbench, 3dmark, * Marius: glmark2 (wayland-es backend works) * https://layers.openembedded.org/layerindex/recipe/5190/ * gfxbench or 3dmark are more featureful tests * there are no recipes * xwayland required * Unigine has downloadable installer, but only x86. Commandline-driven version is not free. * LD_TRACE_LOADED_OBJECTS=1 /usr/bin/foo to find missing libraries * Between Scott, Marius & Jan-Simon check if xwayland can be enabled * Note: talk with upstream wrt enabling xwayland w/o x11 DISTRO feature * 05/03 * could build xwayland (enabled x11 distro-feature), but had to workaround qtwayland issue. * need change in compositor to start xwayland during startup * Mahdi: let's talk about a container build env for devel and ci * Marius: nbd issue: systemd-network-generator kicking in, difference to CI that CI has "systemd.mask=systemd-network-generator.service ip=dhcp" versus "ip=dhcp systemd.mask=systemd-network-generator.service" ... go figure. systemd magic ? * Example of fixed IP setting : * root=/dev/ram0 ramdisk_size=16384 ip=192.168.10.2::192.168.1.1:255.255.0.0::::192.168.1.1 nbd.server=192.168.1.8 systemd.mask=systemd-network-generator.service Next meeting: May 31st !! ==== Meeting 2023/04/19 ==== === Attendees === Jan-Simon, Walt, Scott, Damian, Ishii-san, Yan ===Topics=== * lava lab status * uprev done (server, core-lab, baylibre) * operational * -next branch / Denys * 04/05 * no update * 04/19 * no update * Yocto-check-layer - https://jira.automotivelinux.org/browse/SPEC-4680 * 03/22 * patches to meta-agl-demo merged * 04/19 * meta-agl-devel layers left * Extend Testing * 03/22 * Issue with artifactorial upload, see email. Lava Job 31880. * JS to look * 04/19 * Artifactorial being restarted, including pyartiproxy * IC image and integration of drm-leasemanager * rcar + kvm + pi4 * qemux86-64 * rcar config set to one output on each drm lease * AI1: qemu-x86-64 commandline how to boot the ic-multicontainer image (Damian) * Marius max_outputs=2 == two outputs in qemu * AI2: port above configuration to lava and enable image boot (Jan-Simon) * AI3: extend test to take snapshots within container guests (Jan-Simon and Damian) * AI4: add rcar3 build (kingfisher or refhw tbd) * 02/28 * working on qemu cmdline for multiple/two outputs * 03/22 * Problem: we can setup multiple ports, but they are not connected. Patch sent upstream, but no movement so far. * AI: confirm with Corentin if we can use a dockerized 'custom' qemu * Marius offers to reply with a tested-by: * 04/05 * AI: confirm with Corentin if we can use a dockerized 'custom' qemu * YES: https://lava.baylibre.com/scheduler/job/3473045/definition#defline97 * Option 1: use 1-2 vnc-clients to connect to target and bring screen up - problem needs to connect before cluster starts to connect the output. * Option 2: Custom qemu in docker would already run as patch enables 2nd output by default. * AI(jsmoeller): create qemu with patch and host on docker to test * https://gitlab.com/qemu-project/qemu/-/issues/1107 * WIP version of kernelci/qemu with our patches * AI(Damian): qemu cmdline example (Option2 first , Option1 good for dev) * -device virtio-vga,max_outputs=2,id=vid -display vnc=unix:,display=vid,head=0 -display vnc=unix:,display=vid,head=1 * plus simple client tool to open * Rahul (AMD): benchmarking the OS, gfxbench, 3dmark, * Marius: glmark2 (wayland-es backend works) * https://layers.openembedded.org/layerindex/recipe/5190/ * gfxbench or 3dmark are more featureful tests * there are no recipes * xwayland required * Unigine has downloadable installer, but only x86. Commandline-driven version is not free. * LD_TRACE_LOADED_OBJECTS=1 /usr/bin/foo to find missing libraries * Between Scott, Marius & Jan-Simon check if xwayland can be enabled * Note: talk with upstream wrt enabling xwayland w/o x11 DISTRO feature * Mahdi: let's talk about a container build env for devel and ci Next meeting: May 3rd ==== Meeting 2023/04/05 ==== === Attendees === Jan-Simon, Walt, Damian, Scott, Marius, Victor, Rahul, Du Erpei, Mahdi Shourabi, Kurokawa-san ===Topics=== * lava lab status * uprev done (server, core-lab, baylibre) * operational * -next branch / Denys * 02/28 * fixes required on meta-agl-core for autobuilder to pass * 03/22 * fixes merged * 04/05 * no update * Yocto-check-layer * 02/28 * patches for meta-agl-demo sent to gerrit * 03/22 * patches to meta-agl-demo merged * * Extend Testing * 03/22 * Issue with artifactorial upload, see email. Lava Job 31880. * JS to look * IC image and integration of drm-leasemanager * rcar + kvm + pi4 * qemux86-64 * rcar config set to one output on each drm lease * AI1: qemu-x86-64 commandline how to boot the ic-multicontainer image (Damian) * Marius max_outputs=2 == two outputs in qemu * AI2: port above configuration to lava and enable image boot (Jan-Simon) * AI3: extend test to take snapshots within container guests (Jan-Simon and Damian) * AI4: add rcar3 build (kingfisher or refhw tbd) * 02/28 * working on qemu cmdline for multiple/two outputs * 03/22 * Problem: we can setup multiple ports, but they are not connected. Patch sent upstream, but no movement so far. * AI: confirm with Corentin if we can use a dockerized 'custom' qemu * Marius offers to reply with a tested-by: * 04/05 * AI: confirm with Corentin if we can use a dockerized 'custom' qemu * YES: https://lava.baylibre.com/scheduler/job/3473045/definition#defline97 * Option 1: use 1-2 vnc-clients to connect to target and bring screen up - problem needs to connect before cluster starts to connect the output. * Option 2: Custom qemu in docker would already run as patch enables 2nd output by default. * AI(jsmoeller): create qemu with patch and host on docker to test * AI(Damian): qemu cmdline example (Option2 first , Option1 good for dev) Backlog: * pi4 -> drm issue needs to be reproduced https://lava.automotivelinux.org/scheduler/job/28963#L3744 * 01/31 * issue still present (vc4, drm) * try monitor instead of dongle to rule out dongle * 02/14 * no update * 04/05 * adding 2nd dongle to free port * AI3: POC for binary feed (and what features are included)# * 11/22: * Talk for ALS in preparation (PRServ, HashServ) * 12/13 * Talk slides here: https://static.sched.com/hosted_files/ossjapan2022/84/OSSJ-jsmoeller%20Speed-up%20bitbake.pdf * 01/17 * IC image -> screen snapshot test ? E.g. can disable animation by ENV variable. * 12/13: * Revisit after CES. * 01/17 * Discussing https://gerrit.automotivelinux.org/gerrit/c/AGL/meta-agl-devel/+/28190 * JIRA: SPEC-4606 and SPEC-4608 * Scott will take a look as he had similar issues w/ the KVM image * 01/31 * Multiconfig for IC-container and KVM WIP in support of this (enabling agl-compositor there) * Rahul (AMD): benchmarking the OS, gfxbench, 3dmark, * Marius: glmark2 (wayland-es backend works) * https://layers.openembedded.org/layerindex/recipe/5190/ * gfxbench or 3dmark are more featureful tests * there are no recipes * xwayland required * Unigine has downloadable installer, but only x86. Commandline-driven version is not free. * LD_TRACE_LOADED_OBJECTS=1 /usr/bin/foo to find missing libraries * Between Scott, Marius & Jan-Simon check if xwayland can be enabled * Note: talk with upstream wrt enabling xwayland w/o x11 DISTRO feature * Mahdi: let's talk about a container build env for devel and ci Next meeting: 5th Apr 2023 ==== Meeting 2023/03/22 ==== === Attendees === Jan-Simon, Walt, Marius, Damian, Du Erpei, Mahdi Shourabi, Rahul, Kurokawa-san ! NEW MEETING TIME (day changed to wednesday) ! ===Topics=== * lava lab status * uprev done (server, core-lab, baylibre) * operational * -next branch / Denys * 02/28 * fixes required on meta-agl-core for autobuilder to pass * 03/22 * fixes merged * Yocto-check-layer * 02/28 * patches for meta-agl-demo sent to gerrit * 03/22 * patches to meta-agl-demo merged * Extend Testing * 03/22 * Issue with artifactorial upload, see email. Lava Job 31880. * JS to look * IC image and integration of drm-leasemanager * rcar + kvm + pi4 * qemux86-64 * rcar config set to one output on each drm lease * AI1: qemu-x86-64 commandline how to boot the ic-multicontainer image (Damian) * Marius max_outputs=2 == two outputs in qemu * AI2: port above configuration to lava and enable image boot (Jan-Simon) * AI3: extend test to take snapshots within container guests (Jan-Simon and Damian) * AI4: add rcar3 build (kingfisher or refhw tbd) * 02/28 * working on qemu cmdline for multiple/two outputs * 03/22 * Problem: we can setup multiple ports, but they are not connected. Patch sent upstream, but no movement so far. * AI: confirm with Corentin if we can use a dockerized 'custom' qemu * Marius offers to reply with a tested-by: Backlog: * pi4 -> drm issue needs to be reproduced https://lava.automotivelinux.org/scheduler/job/28963#L3744 * 01/31 * issue still present (vc4, drm) * try monitor instead of dongle to rule out dongle * 02/14 * no update * AI3: POC for binary feed (and what features are included)# * WIP * discussion what the set of packages is that should be in the feed * bitbake world ? * discussion point: qt+html5+flutter demo ? * ask Denys * one idea was to have a minimal bootable image and let the user pull all * need to start somewhere, e.g. agl-demo-platform and expand * Talk about binary feed during ELCE Dublin with chbae * 10/25: * no news * 11/22: * Talk for ALS in preparation (PRServ, HashServ) * 12/13 * Talk slides here: https://static.sched.com/hosted_files/ossjapan2022/84/OSSJ-jsmoeller%20Speed-up%20bitbake.pdf * 01/17 * IC image -> screen snapshot test ? E.g. can disable animation by ENV variable. * 12/13: * Revisit after CES. * 01/17 * Discussing https://gerrit.automotivelinux.org/gerrit/c/AGL/meta-agl-devel/+/28190 * JIRA: SPEC-4606 and SPEC-4608 * Scott will take a look as he had similar issues w/ the KVM image * 01/31 * Multiconfig for IC-container and KVM WIP in support of this (enabling agl-compositor there) New: * Rahul (AMD): benchmarking the OS, gfxbench, 3dmark, * Marius: glmark2 (wayland-es backend works) * https://layers.openembedded.org/layerindex/recipe/5190/ * gfxbench or 3dmark are more featureful tests * there are no recipes * Uniengine has downloadable installer, but only x86. Commandline-driven version is not free. * Next meeting: 5th Apr 2023 ==== Meeting 2023/02/28 ==== === Attendees === Jan-Simon, Walt, Scott, Du Erpei, Damian, Ishii-san, Marius ===Topics=== * lava lab status * operational * PI4 has kernel issues -> graphics does not come up -> see below * preparations for uprev on wednesday (1st march) ~ 3pm UTC * -next branch / Denys * 02/14 * nothing new to report * 02/28 * fixes required on meta-agl-core for autobuilder to pass * Yocto-check-layer * 02/14 * 2 minor issues in layer within meta-agl and meta-agl-demo * continuing with meta-agl-devel * 02/28 * patches for meta-agl-demo sent to gerrit * Extend Testing * 01/17 * Questions by mail * We're debugging a timeout that we see (~30min). We do not know yet what component does cause it. * For now switch to qemux86-64 * 02/14 * Working on talk for AMM * 02/28 * Done * IC image and integration of drm-leasemanager * rcar + kvm + pi4 * qemux86-64 * rcar config set to one output on each drm lease * AI1: qemu-x86-64 commandline how to boot the ic-multicontainer image (Damian) * Marius max_outputs=2 == two outputs in qemu * AI2: port above configuration to lava and enable image boot (Jan-Simon) * AI3: extend test to take snapshots within container guests (Jan-Simon and Damian) * AI4: add rcar3 build (kingfisher or refhw tbd) * 02/28 * working on qemu cmdline for multiple/two outputs Backlog: * pi4 -> drm issue needs to be reproduced https://lava.automotivelinux.org/scheduler/job/28963#L3744 * 01/31 * issue still present (vc4, drm) * try monitor instead of dongle to rule out dongle * 02/14 * no update * AI3: POC for binary feed (and what features are included)# * WIP * discussion what the set of packages is that should be in the feed * bitbake world ? * discussion point: qt+html5+flutter demo ? * ask Denys * one idea was to have a minimal bootable image and let the user pull all * need to start somewhere, e.g. agl-demo-platform and expand * Talk about binary feed during ELCE Dublin with chbae * 10/25: * no news * 11/22: * Talk for ALS in preparation (PRServ, HashServ) * 12/13 * Talk slides here: https://static.sched.com/hosted_files/ossjapan2022/84/OSSJ-jsmoeller%20Speed-up%20bitbake.pdf * 01/17 * IC image -> screen snapshot test ? E.g. can disable animation by ENV variable. * 12/13: * Revisit after CES. * 01/17 * Discussing https://gerrit.automotivelinux.org/gerrit/c/AGL/meta-agl-devel/+/28190 * JIRA: SPEC-4606 and SPEC-4608 * Scott will take a look as he had similar issues w/ the KVM image * 01/31 * Multiconfig for IC-container and KVM WIP in support of this (enabling agl-compositor there) New: * https://gerrit.automotivelinux.org/gerrit/c/AGL/AGL-repo/+/28482 * https://github.com/eclipse/kuksa.val/blob/master/kuksa-val-server/buildinfo.h.in * https://github.com/eclipse/kuksa.val/blob/master/kuksa-val-server/git_watcher.cmake * Next meeting: 21th Mar 2023 ==== Meeting 2023/02/14 ==== === Attendees === Jan-Simon, Walt, Scott, Du Erpei ===Topics=== * lava lab status * operational * PI4 has kernel issues -> graphics does not come up -> see below * lava uprev (replace xnbd-server) * preparations for a mid-feb uprev (after release) - Feb 22-24 would be free * -next branch / Denys * 01/31 * nothing new to report * 02/14 * nothing new to report * Yocto-check-layer * 01/31 * work started * 02/14 * 2 minor issues in layer within meta-agl and meta-agl-demo * continuing with meta-agl-devel * Extend Testing * 01/17 * Questions by mail * We're debugging a timeout that we see (~30min). We do not know yet what component does cause it. * For now switch to qemux86-64 * 02/14 * Working von talk for AMM * pi4 -> drm issue needs to be reproduced https://lava.automotivelinux.org/scheduler/job/28963#L3744 * 01/31 * issue still present (vc4, drm) * try monitor instead of dongle to rule out dongle * 02/14 * no update Backlog: * AI3: POC for binary feed (and what features are included)# * WIP * discussion what the set of packages is that should be in the feed * bitbake world ? * discussion point: qt+html5+flutter demo ? * ask Denys * one idea was to have a minimal bootable image and let the user pull all * need to start somewhere, e.g. agl-demo-platform and expand * Talk about binary feed during ELCE Dublin with chbae * 10/25: * no news * 11/22: * Talk for ALS in preparation (PRServ, HashServ) * 12/13 * Talk slides here: https://static.sched.com/hosted_files/ossjapan2022/84/OSSJ-jsmoeller%20Speed-up%20bitbake.pdf * 01/17 * IC image -> screen snapshot test ? E.g. can disable animation by ENV variable. * 12/13: * Revisit after CES. * 01/17 * Discussing https://gerrit.automotivelinux.org/gerrit/c/AGL/meta-agl-devel/+/28190 * JIRA: SPEC-4606 and SPEC-4608 * Scott will take a look as he had similar issues w/ the KVM image * 01/31 * Multiconfig for IC-container and KVM WIP in support of this (enabling agl-compositor there) New: * IC image and integration of drm-leasemanager * rcar + kvm + pi4 * qemux86-64 * rcar config set to one output on each drm lease * AI1: qemu-x86-64 commandline how to boot the ic-multicontainer image (Damian) * Marius max_outputs=2 == two outputs in qemu * AI2: port above configuration to lava and enable image boot (Jan-Simon) * AI3: extend test to take snapshots within container guests (Jan-Simon and Damian) * AI4: add rcar3 build (kingfisher or refhw tbd) * Next meeting: 28th Feb 2023 ==== Meeting 2023/01/31 ==== === Attendees === Jan-Simon, Walt, Du Erp, Scott, Kurokawa-san, Marius ===Topics=== * lava lab status * operational * pyartiproxy restarted * Issues solved w/ systemd-network-generator, all boards but the pi4 work fine. * PI4 has kernel issues -> graphics does not come up -> see below * lava uprev (replace xnbd-server) * Progress on nbd-server setup * preparations for a mid-feb uprev (after release) * -next branch / Denys * 01/17 * bugfix for YP mickledore * build might be broken with latest upstream * 01/31 * nothing new to report * Yocto-check-layer * 01/31 * work started * Extend Testing * 01/17 * Questions by mail * We're debugging a timeout that we see (~30min). We do not know yet what component does cause it. * For now switch to qemux86-64 * pi4 -> drm issue needs to be reproduced https://lava.automotivelinux.org/scheduler/job/28963#L3744 * 01/31 * issue still present (vc4, drm) * try monitor instead of dongle to rule out dongle Backlog: * AI3: POC for binary feed (and what features are included)# * WIP * discussion what the set of packages is that should be in the feed * bitbake world ? * discussion point: qt+html5+flutter demo ? * ask Denys * one idea was to have a minimal bootable image and let the user pull all * need to start somewhere, e.g. agl-demo-platform and expand * Talk about binary feed during ELCE Dublin with chbae * 10/25: * no news * 11/22: * Talk for ALS in preparation (PRServ, HashServ) * 12/13 * Talk slides here: https://static.sched.com/hosted_files/ossjapan2022/84/OSSJ-jsmoeller%20Speed-up%20bitbake.pdf * 01/17 * IC image -> screen snapshot test ? E.g. can disable animation by ENV variable. * 12/13: * Revisit after CES. * 01/17 * Discussing https://gerrit.automotivelinux.org/gerrit/c/AGL/meta-agl-devel/+/28190 * JIRA: SPEC-4606 and SPEC-4608 * Scott will take a look as he had similar issues w/ the KVM image * 01/31 * Multiconfig for IC-container and KVM WIP in support of this (enabling agl-compositor there) New: * Discuss goals/plans/features for 2023 Next meeting: 14th Feb 2023 ==== Meeting 2023/01/17 ==== === Attendees === Jan-Simon, Walt, Yan, Du Erpei, Marius, Scott, Kurokawa-san ===Topics=== * lava lab status * operational * agl-core-lab resetup * lab partially back * lab-baylibre * refhw throws connection errors * This seems to bite us: * https://github.com/renesas-rcar/meta-renesas/blob/dunfell/meta-rcar-gen3/recipes-core/systemd/systemd-conf_%25.bbappend * adding BBMASK * Marius: * nbd boot issue on pi (master) * check config leaking into image * lava uprev overdue (replace xnbd-server) * Corentin preparing update. Currently checking why new nbd-server is mounted read-only. * https://paste.debian.net/1261480/ * being debugged * * -next branch / Denys * 01/17 * bugfix for YP mickledore * build might be broken with latest upstream * * Extend Testing * 01/17 * Questions by mail * AI3: POC for binary feed (and what features are included)# * WIP * discussion what the set of packages is that should be in the feed * bitbake world ? * discussion point: qt+html5+flutter demo ? * ask Denys * one idea was to have a minimal bootable image and let the user pull all * need to start somewhere, e.g. agl-demo-platform and expand * Talk about binary feed during ELCE Dublin with chbae * 10/25: * no news * 11/22: * Talk for ALS in preparation (PRServ, HashServ) * 12/13 * Talk slides here: https://static.sched.com/hosted_files/ossjapan2022/84/OSSJ-jsmoeller%20Speed-up%20bitbake.pdf * 01/17 * pi4 -> drm issue needs to be reproduced https://lava.automotivelinux.org/scheduler/job/28963#L3744 * 12/13: * No update due to travel * 01/17 * qemuarm -> connman does not ignore network if -> fails due to netboot * 01/17 * Issue solved (adding a !nbdroot match to ignore IF) * That led the connman systemd unit check to fail (due to being ignored). * leads to dns resolution not being up * https://git.automotivelinux.org/AGL/meta-agl/tree/meta-netboot/recipes-core/initramfs-netboot/files/init.sh#n144 * + /usr-merge * check https://git.kernel.org/pub/scm/network/connman/connman.git/tree/ * IC image -> screen snapshot test ? E.g. can disable animation by ENV variable. * 12/13: * Revisit after CES. * 01/17 * Discussing https://gerrit.automotivelinux.org/gerrit/c/AGL/meta-agl-devel/+/28190 * JIRA: SPEC-4606 and SPEC-4608 * Scott will take a look as he had similar issues w/ the KVM image New: * Discuss goals/plans/features for 2023 * Next meeting: 31th January 2023 ==== Meeting 2022/12/13 ==== === Attendees === Jan-Simon, Du Erpei, Marius, Scott ===Topics=== * lava lab status * operational * agl-core-lab resetup * lab partially back * lab-baylibre * refhw throws connection errors * lava uprev overdue (replace xnbd-server) * Corentin preparing update. Currently checking why new nbd-server is mounted read-only. * https://paste.debian.net/1261480/ * being debugged * -next branch / Scott * 11/22 * Scott updated his WIP tree and pushed to gerrit. Sent instructions to Denys to proceed. * YP master uses weston 11 , thus agl-compositor need to be looked at (SPEC-4617) * 12/13 * No update. * Extend Testing * 12/13 * LAVA token issued * use lavacli to setup auth and submit jobs * releng-scripts -> into file , then lavacli can upload job from file * Discussions during the F2F in Berlin show interest in binary feeds. Thus continuing the work on setting up PRServ and Hashserv with readonly ports. Then lateron extending this with a binary feed on master. * AI3: POC for binary feed (and what features are included)# * WIP * discussion what the set of packages is that should be in the feed * bitbake world ? * discussion point: qt+html5+flutter demo ? * ask Denys * one idea was to have a minimal bootable image and let the user pull all * need to start somewhere, e.g. agl-demo-platform and expand * Talk about binary feed during ELCE Dublin with chbae * 10/25: * no news * 11/22: * Talk for ALS in preparation (PRServ, HashServ) * 12/13 * Talk slides here: https://static.sched.com/hosted_files/ossjapan2022/84/OSSJ-jsmoeller%20Speed-up%20bitbake.pdf * pi4 -> drm issue needs to be reproduced https://lava.automotivelinux.org/scheduler/job/28963#L3744 * 10/11 * issue in vc4 driver, later kernel has patch * possibly backport patch * JS to confirm * 11/22: * WIP * 12/13: * No update due to travel * qemuarm -> connman does not ignore network if -> fails due to netboot * 10/25: * only qemuarm does not ignore the network interface where nbdroot is served (we do modify connman configuration file to exclude it in initrd) * 12/13: * No update due to travel * IC image -> screen snapshot test ? E.g. can disable animation by ENV variable. * 11/10 * depends on IC image in CI * 11/22 * Image for IC (and flutter) are in CI now. All are setup to just build on qemux86-64 now. * 12/13: * Revisit after CES. New: * html5 build -> move out of loop and put it at the end (aka won't hold up commit until sstate reuse issue is debugged) * DONE * need to check bbe on needlefish * -> telematic done. Next meeting: 17th January 2023 ==== Meeting 2022/11/22 ==== === Attendees === Jan-Simon, Du Erpei, Yan Xinkuan, Scott, Walt, Marius ===Topics=== * lava lab status * operational * agl-core-lab resetup * lab partially back * lab-baylibre * refhw throws connection errors * lava uprev overdue (replace xnbd-server) * Corentin preparing update. Currently checking why new nbd-server is mounted read-only. * https://paste.debian.net/1261480/ * being debugged * -next branch / Scott * 10/25 * planned for this week * 11/8 * AB issue (weston bbappend) * -> unblock * 11/22 * Scott updated his WIP tree and pushed to gerrit. Sent instructions to Denys to proceed. * YP master uses weston 11 , thus agl-compositor need to be looked at (SPEC-4617) * Extend Testing (Qiu) * 8/30 * Will adopt a ptest-first workflow. Add ptests where necessary and extend only diff to fuego. * rpm ptest: need package for dependency 'fakechroot' (try devtool) . * rpm has ./tests subfolder, can produce pre-compiled rpm-ptest package with it * can work on upstreaming to YP (wrt rpm-ptest and its dependency 'fakechroot') * 9/27 * Working on lava testdefinition (yaml+script) for qa-testdefinitions * 10/11 * we will remove meta-agl/meta-agl-core-test , what is useful can be adopted in meta-agl-devel/meta-agl-test * before adopting or copying, check https://layers.openembedded.org/layerindex/branch/kirkstone/recipes/ is a recipe already exists (e.g. in meta-oe) * 10/25 * meta-agl-core-test removed as discussed * discussion on https://gerrit.automotivelinux.org/gerrit/c/src/agl-test-framework/+/28053 about what to do when ptests are changed/updated/deleted. * 22/8 * Discussion about when to enable the 'crashme' tool. Conclusion: It needs to be selectable (aka user should specifically *enable* the test / possibly expose parameters like length) by the user and not run by default. It also should be the last test to run (due to the crash). * Default inclusion for releng-scripts is defined in these shortcuts. CI user 'ci'(.jinja2) https://git.automotivelinux.org/AGL/releng-scripts/tree/templates/config/default.cfg * * Discussions during the F2F in Berlin show interest in binary feeds. Thus continuing the work on setting up PRServ and Hashserv with readonly ports. Then lateron extending this with a binary feed on master. * AI3: POC for binary feed (and what features are included)# * WIP * discussion what the set of packages is that should be in the feed * bitbake world ? * discussion point: qt+html5+flutter demo ? * ask Denys * one idea was to have a minimal bootable image and let the user pull all * need to start somewhere, e.g. agl-demo-platform and expand * Talk about binary feed during ELCE Dublin with chbae * 10/25: * no news * 11/22: * Talk for ALS in preparation (PRServ, HashServ) * pi4 -> drm issue needs to be reproduced https://lava.automotivelinux.org/scheduler/job/28963#L3744 * 10/11 * issue in vc4 driver, later kernel has patch * possibly backport patch * JS to confirm * 10/25: * no update * 11/22: * WIP * qemuarm -> connman does not ignore network if -> fails due to netboot * 10/25: * only qemuarm does not ignore the network interface where nbdroot is served (we do modify connman configuration file to exclude it in initrd) * IC image -> screen snapshot test ? E.g. can disable animation by ENV variable. * 11/10 * depends on IC image in CI * 11/22 * Image for IC (and flutter) are in CI now. All are setup to just build on qemux86-64 now. New: * html5 build -> move out of loop and put it at the end (aka won't hold up commit until sstate reuse issue is debugged) * need to check bbe on needlefish * Next meeting: 13th December 2022 after that: 17th January ==== Meeting 2022/11/8 ==== === Attendees === Jan-Simon, Du Erpei, Yan Xinkuan, Scott, Walt, Marius ===Topics=== * lava lab status * operational * agl-core-lab resetup * lab partially back * lab-baylibre * refhw throws connection errors * lava uprev overdue (replace xnbd-server) * server need an uprev * time is coordinated between IT and baylibre * -next branch / Scott * 10/25 * planned for this week * 11/8 * AB issue (weston bbappend) * -> unblock * Extend Testing (Qiu) * 8/30 * Will adopt a ptest-first workflow. Add ptests where necessary and extend only diff to fuego. * rpm ptest: need package for dependency 'fakechroot' (try devtool) . * rpm has ./tests subfolder, can produce pre-compiled rpm-ptest package with it * can work on upstreaming to YP (wrt rpm-ptest and its dependency 'fakechroot') * 9/27 * Working on lava testdefinition (yaml+script) for qa-testdefinitions * 10/11 * we will remove meta-agl/meta-agl-core-test , what is useful can be adopted in meta-agl-devel/meta-agl-test * before adopting or copying, check https://layers.openembedded.org/layerindex/branch/kirkstone/recipes/ is a recipe already exists (e.g. in meta-oe) * 10/25 * meta-agl-core-test removed as discussed * discussion on https://gerrit.automotivelinux.org/gerrit/c/src/agl-test-framework/+/28053 about what to do when ptests are changed/updated/deleted. * 11/8 * ==== Meeting 2022/10/25 ==== === Attendees === Jan-Simon, Du Erpei, Scott, Ishii-san, Marius, raouf Riahi, ===Topics=== * lava lab status * operational * agl-core-lab resetup * WIP * lab-baylibre * refhw throws connection errors * lava uprev overdue (replace xnbd-server) * server need an uprev * time is coordinated between IT and baylibre * -next branch / Scott * 9/27 * WIP, works locally, shooting for YP RC timeframe * Likely need to pin meta-clang to pre-v15.0 version * 10/11 * after workshop * 10/25 * planned for this week * Extend Testing (Qiu) * 8/30 * Will adopt a ptest-first workflow. Add ptests where necessary and extend only diff to fuego. * rpm ptest: need package for dependency 'fakechroot' (try devtool) . * rpm has ./tests subfolder, can produce pre-compiled rpm-ptest package with it * can work on upstreaming to YP (wrt rpm-ptest and its dependency 'fakechroot') * 9/27 * Working on lava testdefinition (yaml+script) for qa-testdefinitions * 10/11 * we will remove meta-agl/meta-agl-core-test , what is useful can be adopted in meta-agl-devel/meta-agl-test * before adopting or copying, check https://layers.openembedded.org/layerindex/branch/kirkstone/recipes/ is a recipe already exists (e.g. in meta-oe) * 10/25 * meta-agl-core-test removed as discussed * discussion on https://gerrit.automotivelinux.org/gerrit/c/src/agl-test-framework/+/28053 about what to do when ptests are changed/updated/deleted. * 11/8 * New changes in gerrit https://gerrit.automotivelinux.org/gerrit/c/src/agl-test-framework/+/28119/1 https://gerrit.automotivelinux.org/gerrit/c/src/agl-test-framework/+/28120/1 https://gerrit.automotivelinux.org/gerrit/c/src/agl-test-framework/+/28121/1 * Once https://gerrit.automotivelinux.org/gerrit/c/src/qa-testdefinitions/+/28099 is merged, we can update releng-scripts to call the new testdefinititon. * Discussions during the F2F in Berlin show interest in binary feeds. Thus continuing the work on setting up PRServ and Hashserv with readonly ports. Then lateron extending this with a binary feed on master. * AI3: POC for binary feed (and what features are included)# * WIP * discussion what the set of packages is that should be in the feed * bitbake world ? * discussion point: qt+html5+flutter demo ? * ask Denys * one idea was to have a minimal bootable image and let the user pull all * need to start somewhere, e.g. agl-demo-platform and expand * Talk about binary feed during ELCE Dublin with chbae * 10/25: * no news * pi4 -> drm issue needs to be reproduced https://lava.automotivelinux.org/scheduler/job/28963#L3744 * 10/11 * issue in vc4 driver, later kernel has patch * possibly backport patch * JS to confirm * 10/25: * no update * qemuarm -> connman does not ignore network if -> fails due to netboot * 10/25: * only qemuarm does not ignore the network interface where nbdroot is served (we do modify connman configuration file to exclude it in initrd) * IC image -> screen snapshot test ? E.g. can disable animation by ENV variable. * 11/10 * depends on IC image in CI New: * Backlog: * Damian: adding IC and drm-lease to CI loop (SPEC-4464) * a) adding the IC (agl-lxc) image to jenkins and build the 'lxc-host-image-demo' image. * todo: adding image to CI * b) add drm lease specific test template to qa-testdefinitions (=test scripts) and releng-scripts (template for jenkins to call) * c) then add test in jenkins run specifc to IC/drm-lease * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Also issue in YP upstream Next call: Nov 22nd , same time * Discussions during the F2F in Berlin show interest in binary feeds. Thus continuing the work on setting up PRServ and Hashserv with readonly ports. Then lateron extending this with a binary feed on master. * AI3: POC for binary feed (and what features are included)# * WIP * discussion what the set of packages is that should be in the feed * bitbake world ? * discussion point: qt+html5+flutter demo ? * ask Denys * one idea was to have a minimal bootable image and let the user pull all * need to start somewhere, e.g. agl-demo-platform and expand * Talk about binary feed during ELCE Dublin with chbae * 10/25: * no news * pi4 -> drm issue needs to be reproduced https://lava.automotivelinux.org/scheduler/job/28963#L3744 * 10/11 * issue in vc4 driver, later kernel has patch * possibly backport patch * JS to confirm * 10/25: * no update * qemuarm -> connman does not ignore network if -> fails due to netboot * 10/25: * only qemuarm does not ignore the network interface where nbdroot is served (we do modify connman configuration file to exclude it in initrd) * IC image -> screen snapshot test ? E.g. can disable animation by ENV variable. * 11/10 * depends on IC image in CI New: * Backlog: * Damian: adding IC and drm-lease to CI loop (SPEC-4464) * a) adding the IC (agl-lxc) image to jenkins and build the 'lxc-host-image-demo' image. * todo: adding image to CI * b) add drm lease specific test template to qa-testdefinitions (=test scripts) and releng-scripts (template for jenkins to call) * c) then add test in jenkins run specifc to IC/drm-lease * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Also issue in YP upstream Next call: Nov 22th , same time ==== Meeting 2022/10/25 ==== === Attendees === Jan-Simon, Du Erpei, Scott, Ishii-san, Marius, raouf Riahi, ===Topics=== * lava lab status * operational * agl-core-lab resetup * WIP * lab-baylibre * refhw throws connection errors * lava uprev overdue (replace xnbd-server) * server need an uprev * time is coordinated between IT and baylibre * -next branch / Scott * 9/27 * WIP, works locally, shooting for YP RC timeframe * Likely need to pin meta-clang to pre-v15.0 version * 10/11 * after workshop * 10/25 * planned for this week * Extend Testing (Qiu) * 8/30 * Will adopt a ptest-first workflow. Add ptests where necessary and extend only diff to fuego. * rpm ptest: need package for dependency 'fakechroot' (try devtool) . * rpm has ./tests subfolder, can produce pre-compiled rpm-ptest package with it * can work on upstreaming to YP (wrt rpm-ptest and its dependency 'fakechroot') * 9/27 * Working on lava testdefinition (yaml+script) for qa-testdefinitions * 10/11 * we will remove meta-agl/meta-agl-core-test , what is useful can be adopted in meta-agl-devel/meta-agl-test * before adopting or copying, check https://layers.openembedded.org/layerindex/branch/kirkstone/recipes/ is a recipe already exists (e.g. in meta-oe) * 10/25 * meta-agl-core-test removed as discussed * discussion on https://gerrit.automotivelinux.org/gerrit/c/src/agl-test-framework/+/28053 about what to do when ptests are changed/updated/deleted. * Discussions during the F2F in Berlin show interest in binary feeds. Thus continuing the work on setting up PRServ and Hashserv with readonly ports. Then lateron extending this with a binary feed on master. * AI3: POC for binary feed (and what features are included)# * WIP * discussion what the set of packages is that should be in the feed * bitbake world ? * discussion point: qt+html5+flutter demo ? * ask Denys * one idea was to have a minimal bootable image and let the user pull all * need to start somewhere, e.g. agl-demo-platform and expand * Talk about binary feed during ELCE Dublin with chbae * 10/25: * no news * pi4 -> drm issue needs to be reproduced https://lava.automotivelinux.org/scheduler/job/28963#L3744 * 10/11 * issue in vc4 driver, later kernel has patch * possibly backport patch * JS to confirm * 10/25: * no update * qemuarm -> connman does not ignore network if -> fails due to netboot * 10/25: * only qemuarm does not ignore the network interface where nbdroot is served (we do modify connman configuration file to exclude it in initrd) * IC image -> screen snapshot test ? E.g. can disable animation by ENV variable. * 11/10 * depends on IC image in CI New: * Backlog: * Damian: adding IC and drm-lease to CI loop (SPEC-4464) * a) adding the IC (agl-lxc) image to jenkins and build the 'lxc-host-image-demo' image. * todo: adding image to CI * b) add drm lease specific test template to qa-testdefinitions (=test scripts) and releng-scripts (template for jenkins to call) * c) then add test in jenkins run specifc to IC/drm-lease * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Also issue in YP upstream Next call: Nov 8th , same time ==== Meeting 2022/10/11 ==== === Attendees === Jan-Simon, Du Erpei, Scott, Ishii-san, Marius ===Topics=== * lava lab status * operational * agl-core-lab resetup * WIP * lab-baylibre * refhw throws connection errors * lava uprev overdue (replace xnbd-server) * -next branch / Scott * 9/27 * WIP, works locally, shooting for YP RC timeframe * Likely need to pin meta-clang to pre-v15.0 version * 10/11 * after workshop * Extend Testing (Qiu) * 8/30 * Will adopt a ptest-first workflow. Add ptests where necessary and extend only diff to fuego. * rpm ptest: need package for dependency 'fakechroot' (try devtool) . * rpm has ./tests subfolder, can produce pre-compiled rpm-ptest package with it * can work on upstreaming to YP (wrt rpm-ptest and its dependency 'fakechroot') * 9/27 * Working on lava testdefinition (yaml+script) for qa-testdefinitions * 10/11 * we will remove meta-agl/meta-agl-core-test , what is useful can be adopted in meta-agl-devel/meta-agl-test * before adopting or copying, check https://layers.openembedded.org/layerindex/branch/kirkstone/recipes/ is a recipe already exists (e.g. in meta-oe) * Discussions during the F2F in Berlin show interest in binary feeds. Thus continuing the work on setting up PRServ and Hashserv with readonly ports. Then lateron extending this with a binary feed on master. * AI3: POC for binary feed (and what features are included)# * WIP * discussion what the set of packages is that should be in the feed * bitbake world ? * discussion point: qt+html5+flutter demo ? * ask Denys * one idea was to have a minimal bootable image and let the user pull all * need to start somewhere, e.g. agl-demo-platform and expand * Talk about binary feed during ELCE Dublin with chbae * pi4 -> drm issue needs to be reproduced https://lava.automotivelinux.org/scheduler/job/28963#L3744 * 11/10 * issue in vc4 driver, later kernel has patch * possibly backport patch * JS to confirm * qemuarm -> connman does not ignore network if -> fails due to netboot * IC image -> screen snapshot test ? E.g. can disable animation by ENV variable. * 11/10 * depends on IC image in CI New: * Backlog: * Damian: adding IC and drm-lease to CI loop (SPEC-4464) * a) adding the IC (agl-lxc) image to jenkins and build the 'lxc-host-image-demo' image. * todo: adding image to CI * b) add drm lease specific test template to qa-testdefinitions (=test scripts) and releng-scripts (template for jenkins to call) * c) then add test in jenkins run specifc to IC/drm-lease * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Also issue in YP upstream Next call: Oct 11th , same time -------- ==== Meeting 2022/09/27 ==== === Attendees === Jan-Simon, Du Erpei, Scott, Marius ===Topics=== * lava lab status * operational * agl-core-lab resetup * WIP * lab-baylibre * operational * lava uprev overdue (replace xnbd-server) * -next branch / Scott * 9/27 * WIP, works locally, shooting for YP RC timeframe * Likely need to pin meta-clang to pre-v15.0 version * Extend Testing (Qiu) * 8/30 * Will adopt a ptest-first workflow. Add ptests where necessary and extend only diff to fuego. * rpm ptest: need package for dependency 'fakechroot' (try devtool) . * rpm has ./tests subfolder, can produce pre-compiled rpm-ptest package with it * can work on upstreaming to YP (wrt rpm-ptest and its dependency 'fakechroot') * 9/27 * Working on lava testdefinition (yaml+script) for qa-testdefinitions * Discussions during the F2F in Berlin show interest in binary feeds. Thus continuing the work on setting up PRServ and Hashserv with readonly ports. Then lateron extending this with a binary feed on master. * AI3: POC for binary feed (and what features are included)# * WIP * discussion what the set of packages is that should be in the feed * bitbake world ? * discussion point: qt+html5+flutter demo ? * ask Denys * one idea was to have a minimal bootable image and let the user pull all * need to start somewhere, e.g. agl-demo-platform and expand * Talk about binary feed during ELCE Dublin with chbae New: * pi4 -> drm issue needs to be reproduced https://lava.automotivelinux.org/scheduler/job/28963#L3744 * watch out for connman.service failures * IC image -> screen snapshot test ? E.g. can disable animation by ENV variable. * Discussed UI snapshot for the different variants (IVI, HTML5, flutter, IC) and how to enable it. Backlog: * Damian: adding IC and drm-lease to CI loop (SPEC-4464) * a) adding the IC (agl-lxc) image to jenkins and build the 'lxc-host-image-demo' image. * todo: adding image to CI * b) add drm lease specific test template to qa-testdefinitions (=test scripts) and releng-scripts (template for jenkins to call) * c) then add test in jenkins run specifc to IC/drm-lease * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Also issue in YP upstream Next call: Oct 11th , same time ==== Meeting 2022/08/30 ==== === Attendees === Jan-Simon, Qiu, Du Erpei, Scott, Marius ===Topics=== * lava lab status * agl-core-lab resetup * WIP * lab-baylibre * operational * -next branch / Scott * test builds up to homescreen, waiting for fixes on warnings upstream * WIP but close * Extend Testing (Qiu) * 8/2 * above changes got merged * next steps: * need to enable feature - target is agl-(ivi-)demo-platform * need a lava test definition to call it * then adding more test cases / test sets * 8/30 * Will adopt a ptest-first workflow. Add ptests where necessary and extend only diff to fuego. * rpm ptest: need package for dependency 'fakechroot' (try devtool) . * rpm has ./tests subfolder, can produce pre-compiled rpm-ptest package with it * can work on upstreaming to YP (wrt rpm-ptest and its dependency 'fakechroot') * Discussions during the F2F in Berlin show interest in binary feeds. Thus continuing the work on setting up PRServ and Hashserv with readonly ports. Then lateron extending this with a binary feed on master. * AI1: setup new VM for PRServ/new and Hashserv/new. * done * AI2: change CI in master to use it * done * AI3: POC for binary feed (and what features are included)# * WIP * discussion what the set of packages is that should be in the feed * bitbake world ? * discussion point: qt+html5+flutter demo ? * ask Denys * one idea was to have a minimal bootable image and let the user pull all * need to start somewhere, e.g. agl-demo-platform and expand * Damian: adding IC and drm-lease to CI loop (SPEC-4464) * a) adding the IC (agl-lxc) image to jenkins and build the 'lxc-host-image-demo' image. * todo: adding image to CI * b) add drm lease specific test template to qa-testdefinitions (=test scripts) and releng-scripts (template for jenkins to call) * c) then add test in jenkins run specifc to IC/drm-lease New: * IC image -> screen snapshot test ? E.g. can disable animation by ENV variable. * Discussed UI snapshot for the different variants (IVI, HTML5, flutter, IC) and how to enable it. Backlog: * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Also issue in YP upstream Next call: Sep 20th , same time ==== Meeting 2022/08/02 ==== === Attendees === Jan-Simon, Du Erpei, Kurokawa-san, Scott, Marius, Ishii-san ===Topics=== * lava lab status * agl-core-lab resetup * working on worker/dispatcher node * working on h3ulcb+kf * lab-baylibre * had planned power outage * systems coming back * -next branch / Scott * will try next week * Extend Testing (Qiu) * Will follow-up by jira and email. * 5/25 will use src/qa-test-misc for testframework sources * want to work on nightly and release tests * 6/7 * WIP * 7/5 * https://jira.automotivelinux.org/browse/SPEC-4345 * Presented * Will comment in jira SPEC-4345 * 7/19 * reviewing https://gerrit.automotivelinux.org/gerrit/c/src/agl-test-framework/+/27697/ * and https://gerrit.automotivelinux.org/gerrit/c/src/qa-test-misc/+/27698 * 8/2 * above changes got merged * next steps: * need to enable feature - target is agl-(ivi-)demo-platform * need a lava test definition to call it * then adding more test cases / test sets * Discussions during the F2F in Berlin show interest in binary feeds. Thus continuing the work on setting up PRServ and Hashserv with readonly ports. Then lateron extending this with a binary feed on master. * AI1: setup new VM for PRServ/new and Hashserv/new. * 7/5: * New servers are up and change to use them is in gerrit. * 7/19: * prserv r/o is on: prserv.automotivelinux.org:8282 * hashserv r/o is on: hashequiv.automotivelinux.org:8484 * * AI2: change CI in master to use it * 7/5: * https://gerrit.automotivelinux.org/gerrit/c/AGL/meta-agl/+/27699 * done * AI3: POC for binary feed (and what features are included)# * WIP * discussion what the set of packages is that should be in the feed * bitbake world ? * discussion point: qt+html5+flutter demo ? * ask Denys * one idea was to have a minimal bootable image and let the user pull all * need to start somewhere, e.g. agl-demo-platform and expand * Damian: adding IC and drm-lease to CI loop (SPEC-4464) * a) adding the IC (agl-lxc) image to jenkins and build the 'lxc-host-image-demo' image. * 8/2: jsmoeller: WIP * 8/2: can move meta-agl-lxc --> meta-agl-ic-container now * * b) add drm lease specific test template to qa-testdefinitions (=test scripts) and releng-scripts (template for jenkins to call) * c) then add test in jenkins run specifc to IC/drm-lease New: * IC image -> screen snapshot test ? E.g. can disable animation by ENV variable. * Discussed UI snapshot for the different variants (IVI, HTML5, flutter, IC) and how to enable it. Backlog: * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Also issue in YP upstream Next call: Aug 30th , same time ==== Meeting 2022/07/19 ==== === Attendees === Jan-Simon, Walt, Du Erpei, Damian Hobson-Garcia, Kurokawa-san, Scott, Marius ===Topics=== * lava lab status * agl-core-lab down (due to move) * lab-baylibre operational * 17/07 * fixed u-boot load addr * systemd configure the network - systemd.mask=systemd-networkd.service * for investigation wrt systemd-networkd * -next branch / Scott * after EW * likely in 2 weeks * Extend Testing (Qiu) * Will follow-up by jira and email. * 5/25 will use src/qa-test-misc for testframework sources * want to work on nightly and release tests * 6/7 * WIP * 7/5 * https://jira.automotivelinux.org/browse/SPEC-4345 * Presented * Will comment in jira SPEC-4345 * 7/19 * reviewing https://gerrit.automotivelinux.org/gerrit/c/src/agl-test-framework/+/27697/ * and https://gerrit.automotivelinux.org/gerrit/c/src/qa-test-misc/+/27698 * Discussions during the F2F in Berlin show interest in binary feeds. Thus continuing the work on setting up PRServ and Hashserv with readonly ports. Then lateron extending this with a binary feed on master. * AI1: setup new VM for PRServ/new and Hashserv/new. * 7/5: * New servers are up and change to use them is in gerrit. * 7/19: * prserv r/o is on: prserv.automotivelinux.org:8282 * hashserv r/o is on: hashequiv.automotivelinux.org:8484 * * AI2: change CI in master to use it * 7/5: * https://gerrit.automotivelinux.org/gerrit/c/AGL/meta-agl/+/27699 * done * AI3: POC for binary feed (and what features are included)# New: * Damian: adding IC and drm-lease to CI loop (SPEC-4464) * a) adding the IC (agl-lxc) image to jenkins and build the 'lxc-host-image-demo' image. * b) add drm lease specific test template to qa-testdefinitions (=test scripts) and releng-scripts (template for jenkins to call) * c) then add test in jenkins run specifc to IC/drm-lease Backlog: * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Also issue in YP upstream Next call: Aug 2nd , same time ==== Meeting 2022/07/05 ==== === Attendees === Jan-Simon, Walt, Du Erpei, Qiu Tingting, Marius ===Topics=== * lava lab status * agl-core-lab down (due to move) * lab-baylibre operational * 06/07 * ndb stalls currently, investigation ongoing * -next branch / Scott * after EW * likely in 2 weeks * Extend Testing (Qiu) * Will follow-up by jira and email. * 5/25 will use src/qa-test-misc for testframework sources * want to work on nightly and release tests * 6/7 * WIP * 7/5 * https://jira.automotivelinux.org/browse/SPEC-4345 * Presented * Will comment in jira SPEC-4345 * * Discussions during the F2F in Berlin show interest in binary feeds. Thus continuing the work on setting up PRServ and Hashserv with readonly ports. Then lateron extending this with a binary feed on master. * AI1: setup new VM for PRServ/new and Hashserv/new. * 7/5: * New servers are up and change to use them is in gerrit. * AI2: change CI in master to use it * 7/5: * https://gerrit.automotivelinux.org/gerrit/c/AGL/meta-agl/+/27699 * AI3: POC for binary feed (and what features are included)# * Change in master to remove ${AUTOREV} from app builds reduced build times by 40% * Question if we remove ${AUTOREV} in marlin and/or lamprey as well New: * basesystem got disabled as requested by Date-san * this unblocks meta-agl-devel * BBE was disabled (missing kirkstone support) * meta-agl in gerrit adds additional info on commit - compare: * https://git.automotivelinux.org/AGL/meta-agl/commit/?id=34309bc1e6b092e3af5c5d559ad17cee77e99eca * https://git.automotivelinux.org/AGL/meta-agl-demo/commit/?id=eb0faffd70335e38a51ea153f75b6a2c4ac3e3c5 * Backlog: * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Also issue in YP upstream Next call: Jul 19th , same time ==== Meeting 2022/06/07 ==== === Attendees === Jan-Simon, Kurokawa-san, Marius, Scott, Duerp, Ishii-san ===Topics=== * lava lab status * lava master down with 500, filed bugreport w/ IT * resolved * agl-core-lab down (due to move) * lab-baylibre operational * 06/07 * ndb stalls currently , will work with Corentin to resolve * -next branch / Scott * after EW * Extend Testing (Qiu) * Will follow-up by jira and email. * 5/25 will use src/qa-test-misc for testframework sources * want to work on nightly and release tests * 6/7 * WIP * Discussions during the F2F in Berlin show interest in binary feeds. Thus continuing the work on setting up PRServ and Hashserv with readonly ports. Then lateron extending this with a binary feed on master. * AI1: setup new VM for PRServ/new and Hashserv/new. * AI2: change CI in master to use it * AI3: POC for binary feed (and what features are included)# New: * basesystem got disabled as requested by Date-san * this unblocks meta-agl-devel * BBE was disabled (missing kirkstone support) * meta-agl in gerrit adds additional info on commit - compare: * https://git.automotivelinux.org/AGL/meta-agl/commit/?id=34309bc1e6b092e3af5c5d559ad17cee77e99eca * https://git.automotivelinux.org/AGL/meta-agl-demo/commit/?id=eb0faffd70335e38a51ea153f75b6a2c4ac3e3c5 * Backlog: * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Also issue in YP upstream Next call: Jul 5th , same time ==== Meeting 2022/05/24 ==== === Attendees === Jan-Simon, Walt, Scott, Marius, Denys ===Topics=== * lava lab status * lava master down with 500, filed bugreport w/ IT * agl-core-lab down (due to move) * lab-baylibre operational * -next branch / Scott * 5/25: merged to master, will look at YP langdale -next soon * Ned talking to Renesas about V3H * Extend Testing (Qiu) * Will follow-up by jira and email. * 5/25 will use src/qa-test-misc for testframework sources * want to work on nightly and release tests New: * Discussions during the F2F in Berlin show interest in binary feeds. Thus continuing the work on setting up PRServ and Hashserv with readonly ports. Then lateron extending this with a binary feed on master. * AI1: setup new VM for PRServ/new and Hashserv/new. * AI2: change CI in master to use it * AI3: POC for binary feed (and what features are included)# * pi4 / screenshot issues * https://jira.automotivelinux.org/browse/SPEC-4379 Backlog: * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Also issue in YP upstream Next call: May 10th , same time ==== Meeting 2022/04/26 ==== === Attendees === Jan-Simon, Marius, Kurokawa-san, Duerpei ===Topics=== * lava lab status * agl-core-lab down (due to move) * lab-baylibre operational * planned lava update in march after marlin release * 4/26: update being prepared by Corentin * h3ulcb (h3ulcb-nogfx image) - resolution issue on master still an issue * u-boot loadaddr checked, next: attach display instead of jig, * 03/15: wip, no update * SPEC-4275 (HDMI-jig, works with Gechic), SPEC-4276 bad page messages on h3ulcb * maybe write custom dtb to supply memory to h3 instead of autodetection ? * firmware flashing via cpld as documented on elinux.org * 04/26: reflashed uboot and adapted loadaddr * -next branch / Scott * 03/15 * python cryptography module in oe-core now * rust is becoming a dependency (e.g. gstreamer) * mesa 22 probably as well * refresh next week, then see if branch is available upstream or if we pick a revision for the initial uprev of master. * 3/29 * Some updates last week, will push manifest udpate. * experiments with usrmerge * no honister branch, yet * flutter and html do not build. issue is with 'gn'. * lets ask YP for timeframe of branch * 4/26 * Being merged to master * Ronan has a sandbox with more changes specific to R-Car * agl-lxc has recipe issues, will need updates * * Ned talking to Renesas about V3H * Extend Testing (Qiu) * Will follow-up by jira and email. New: * Backlog: * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Discussion on required changes for CI * next spring 3.5 prserv/hashserv (rw+ro) Next call: May 10th , same time ==== Meeting 2022/03/08 ==== === Attendees === Jan-Simon, Walt, Scott, marius, Duerpei ===Topics=== * lava lab status * operational * planned lava update in march after marlin release * h3ulcb (h3ulcb-nogfx image) - resolution issue on master still an issue * u-boot loadaddr checked, next: attach display instead of jig, * 03/15: wip, no update * SPEC-4275 (HDMI-jig, works with Gechic), SPEC-4276 bad page messages on h3ulcb * maybe write custom dtb to supply memory to h3 instead of autodetection ? * firmware flashing via cpld as documented on elinux.org * -next branch / Scott * 03/15 * python cryptography module in oe-core now * rust is becoming a dependency (e.g. gstreamer) * mesa 22 probably as well * refresh next week, then see if branch is available upstream or if we pick a revision for the initial uprev of master. * 3/29 * Some updates last week, will push manifest udpate. * experiments with usrmerge * no honister branch, yet * flutter and html do not build. issue is with 'gn'. * lets ask YP for timeframe of branch * When to switch master to kirkstone * April timeframe right after upstream release * Upstream seem to be on track * Ned talking to Renesas about V3H * reviewed Task list for SAT / SC New: * SDK * Backlog: * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Discussion on required changes for CI * next spring 3.5 prserv/hashserv (rw+ro) Next call: Apr 12th, same time ==== Meeting 2022/03/08 ==== === Attendees === Jan-Simon, Walt, Scott, Du Rrpei ===Topics=== * lava lab status * operational * planned lava update in march after marlin release * h3ulcb (h3ulcb-nogfx image) - resolution issue on master still an issue * u-boot loadaddr checked, next: attach display instead of jig, * 03/15: wip, no update * -next branch / Scott * 03/08 * license now SPDX-conform, there is a script to convert * all layers bumped * compositor not yet rebased * html5 image test possibly next week * python build changes (wheels) don't affect us atm * need to watch meta-qt5 (and possibly need to backport patches from KDE) * https://community.kde.org/Qt5PatchCollection * ssl updates needed for nav app * 03/15 * python cryptography module in oe-core now * rust is becoming a dependency (e.g. gstreamer) * mesa 22 probably as well * refresh next week, then see if branch is available upstream or if we pick a revision for the initial uprev of master. * When to switch master to kirkstone * April timeframe right after upstream release * Upstream seem to be on track * Ned talking to Renesas about V3H * reviewed Task list for SAT / SC New: * Backlog: * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Discussion on required changes for CI * next spring 3.5 prserv/hashserv (rw+ro) Next call: Mar 29st, same time ==== Meeting 2022/03/08 ==== === Attendees === Jan-Simon, Walt, Scott, Marius ===Topics=== * lava lab status * operational * planned lava update in march after marlin release * am335x-sancloud-bbe-02 shows a kernel fault in some jobs, the other bbe does not. * traced down to CONFIG_RELAY in the linux kernel. disabled for bbe * h3ulcb (h3ulcb-nogfx image) - resolution issue on master still an issue * u-boot loadaddr checked, next: attach display instead of jig, * -next branch / Scott * 02/1 * rebase pushed last monday, 3.5m2 is next target * Inclusive Language renaming , might affect us in Kirkstone or on autobuilder * 02/15 * glibc update in upstream YP, a mesa update might land as well * 03/08 * license now SPDX-conform, there is a script to convert * all layers bumped * compositor not yet rebased * html5 image test possibly next week * python build changes (wheels) don't affect us atm * need to watch meta-qt5 (and possibly need to backport patches from KDE) * https://community.kde.org/Qt5PatchCollection * ssl updates needed for nav app * Can bring back agl-compositor screenshot test * 02/01 * qemuarm64/qemuarm : screenshooter does not work, next: JS: live instance debug w/marius. * 03/08 * fixed. had to reload systemd daemon in user session as well to make New: * When to switch master to kirkstone * April timeframe right after upstream release * Ned talking to Renesas about V3H * reviewed Task list for SAT / SC Backlog: * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Discussion on required changes for CI * next spring 3.5 prserv/hashserv (rw+ro) Next call: Mar 15st, same time ==== Meeting 2022/02/15 ==== === Attendees === Jan-Simon, Walt, Scott, HeeWon Kim (DrimAES), Ned, ===Topics=== * lava lab status * operational * planned lava update in march after marlin release * am335x-sancloud-bbe-02 shows a kernel fault in some jobs, the other bbe does not. * -next branch / Scott * 2/1 * rebase pushed last monday, 3.5m2 is next target * Inclusive Language renaming , might affect us in Kirkstone or on autobuilder * 2/15 * glibc update in upstream YP, a mesa update might land as well * Can bring back agl-compositor screenshot test * 02/01 * qemuarm64/qemuarm : screenshooter does not work, next: JS: live instance debug w/marius. New: * MM m2 release with html5 and flutter builds * When to switch master to kirkstone * April timeframe right after upstream release * Ned talking to Renesas about V3H Backlog: * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Discussion on required changes for CI * next spring 3.5 prserv/hashserv (rw+ro) Next call: Mar 1st, same time ==== Meeting 2022/02/01 ==== === Attendees === Jan-Simon, Walt, Scott, Marius ===Topics=== * lava lab status * operational * planned lava update in march after marlin release * am335x-sancloud-bbe-02 shows a kernel fault in some jobs, the other bbe does not. * -next branch / Scott * 1/18 * rebase done last week, looks fine. * (no)network namespace landed in bitbake * AI: check CI distro support and if necessary prepare uprev (ubuntu 20.04 or newer) * 2/1 * rebase pushed last monday, 3.5m2 is next target * Inclusive Language renaming , might affect us in Kirkstone or on autobuilder * Can bring back agl-compositor screenshot test * 01/18 * talking about https://lava.automotivelinux.org/scheduler/job/23861 , need to check for a timeout (e.g. systemd unit) * 02/01 * qemuarm64/qemuarm : screenshooter does not work, next: JS: live instance debug w/marius. New: * Backlog: * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Discussion on required changes for CI * next spring 3.5 prserv/hashserv (rw+ro) Next call: Feb 15th, same time === Attendees === Jan-Simon, Walt, Qiu Tingting, Scott, Marius ===Topics=== * lava lab status * operational * -next branch / Scott * 12/07 * New Readme file check upstream (yocto-check-layer) * "Maintainer" * "Patch" * -> Jira and upreving master * Upstream might warn about upstream-status and 'Pending' state * https://git.yoctoproject.org/poky/tree/scripts/contrib/patchreview.py alike * 1/18 * rebase done last week, looks fine. * (no)network namespace landed in bitbake * AI: check CI distro support and if necessary prepare uprev (ubuntu 20.04 or newer) * Can bring back agl-compositor screenshot test * 12/07: * added back, working on issue on ARM (32+64) SPEC-4165 * discussed possible issues * recent jobs using same env: * lamprey: https://lava.automotivelinux.org/scheduler/job/22984 * master: https://lava.automotivelinux.org/scheduler/job/23038 * 01/18 * talking about https://lava.automotivelinux.org/scheduler/job/23861 , need to check for a timeout (e.g. systemd unit) * memory check - m3/h3 had only 4G after bsp update (firmware/bios/uboot older + newer kernel) * h3+kf and refhw could check for 8GB * https://jira.automotivelinux.org/browse/SPEC-4169 * Marius will verify if it still applies and update the jira. New: * Marius: * camera and videocore firmware * https://www.raspberrypi.com/documentation/computers/config_txt.html#boot-options * https://git.automotivelinux.org/AGL/meta-agl/tree/meta-agl-bsp/meta-raspberrypi/recipes-bsp/bootfiles/rpi-config_git.bbappend * Backlog: * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Discussion on required changes for CI * jenkins-job-builder vs gitlab-runner exec ? * next spring 3.5 prserv/hashserv (rw+ro) Next call: Feb 1st, same time ==== Meeting 2021/12/07 ==== === Attendees === Jan-Simon, Scott, Marius, Qiu, Ishii-san ===Topics=== * lava lab status * operational * lava-docker * 12/07 h3ulcb+kf in baylibre lab reflashed * -next branch / Scott * 12/07 * New Readme file check upstream (yocto-check-layer) * "Maintainer" * "Patch" * -> Jira and upreving master * Upstream might warn about upstream-status and 'Pending' state * https://git.yoctoproject.org/poky/tree/scripts/contrib/patchreview.py alike * Can bring back agl-compositor screenshot test * need to write /etc/default/homescreen HOMESCREEN_DEMO_CI=1 * Fix: https://git.automotivelinux.org/src/qa-testdefinitions/tree/common/scripts/agl-screenshot- * systemctl restart weston@display.service --> weston.service * compositor.log needs to come from journal * 12/07: * added back, working on issue on ARM (32+64) SPEC-4165 * discussed possible issues * recent jobs using same env: * lamprey: https://lava.automotivelinux.org/scheduler/job/22984 * master: https://lava.automotivelinux.org/scheduler/job/23038 * memory check - m3/h3 had only 4G after bsp update (firmware/bios/uboot older + newer kernel) * h3+kf and refhw could check for 8GB New: * https://jira.automotivelinux.org/browse/SPEC-4169 * Jan-Simon will check the SDK artifacts * Marius: * camera and videocore firmware * https://www.raspberrypi.com/documentation/computers/config_txt.html#boot-options * https://git.automotivelinux.org/AGL/meta-agl/tree/meta-agl-bsp/meta-raspberrypi/recipes-bsp/bootfiles/rpi-config_git.bbappend * Backlog: * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Discussion on required changes for CI * jenkins-job-builder vs gitlab-runner exec ? * next spring 3.5 prserv/hashserv (rw+ro) Next call: Dec 21th, same time ==== Meeting 2021/11/23 ==== === Attendees === Jan-Simon, Walt, Scott, Marius, Qiu ===Topics=== * lava lab status * lava-docker * 11/03 master conversion caused a few red builds. issues are fixed. * 11/26 h3ulcb+kf in baylibre lab to be reflashed * -next branch / Scott * 11/03 * appfw removal rebased and pushed, removed a couple of commits * 11/26 * rebase done, looking at pi4 and etnaviv as time permits * Discussion on required changes for CI * jenkins-job-builder vs gitlab-runner exec ? * next spring 3.5 prserv/hashserv (rw+ro) * Can bring back agl-compositor screenshot test * need to write /etc/default/homescreen HOMESCREEN_DEMO_CI=1 * Fix: https://git.automotivelinux.org/src/qa-testdefinitions/tree/common/scripts/agl-screenshot- * systemctl restart weston@display.service --> weston.service * compositor.log needs to come from journal * 11/26: * working on script. agl-compositor from journal * memory check - m3/h3 had only 4G after bsp update (firmware/bios/uboot older + newer kernel) WIP: * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ New: Next call: Dec 7th, same time ==== Meeting 2021/11/03 ==== === Attendees === Jan-Simon, Walt, Scott, Kurokawa-san, Alperen Bulut, Marius, Raouf Riahi, Ishii-san ===Topics=== * lava lab status * lava-docker * 11/03 all operational * 11/03 master conversion caused a few red builds. issues are fixed. * -next branch / Scott * 11/03 * appfw removal rebased and pushed, removed a couple of commits * Discussion on required changes for CI * jenkins-job-builder vs gitlab-runner exec ? * next spring 3.5 prserv/hashserv (rw+ro) * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * 11/03 New: * Can bring back agl-compositor screenshot test * need to write /etc/default/homescreen HOMESCREEN_DEMO_CI=1 * Fix: https://git.automotivelinux.org/src/qa-testdefinitions/tree/common/scripts/agl-screenshot- * systemctl restart weston@display.service --> weston.service * compositor.log needs to come from journal * memory check - m3/h3 had only 4G after bsp update (firmware/bios/uboot older + newer kernel) Next call: Nov 23, same time ==== Meeting 2021/10/11 ==== === Attendees === Jan-Simon, Scott, Li Xiaoming, Kurokawa-san ===Topics=== * lava lab status * lava-docker * 10/26 all operational * pyagl: * few changes in flight , should review+merge asap + check backport to lamprey * ^^ before removing appfw. * -next branch / Scott * 10/12 * qtwayland patch disabled in -next , IC Group needs to sort out * sandbox or snapshot for when honister is branched (for reference) * 10/26 * sandbox/scottm/next-honister capturing the state of 3.4-rc1 for reference uploaded * noupdates planned atm beside honister final * updated next branch to latest master (python 3.10 et al, large changes in upstream) * meta-agl-core works but lua in poky is newer than wireplumber expects * Javier: update on openQA+lava - got POC working * 8/31 * WIP * 9/14 * no update * 10/12 * -> Presentation @ELC and upcoming ALS * Discussion on required changes for CI * build targets * agl-demo-platform(-crosssdk) - alias agl-image-ivi or agl-image-weston * lava tests to change * agl-compositor/screenshooter: minimal change around systemd unit * remove pyagl * remove application-lifecycle * jenkins-job-builder vs gitlab-runner exec ? * next spring 3.5 prserv/hashserv (rw+ro) * "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * 10/12: vacation * 10/26: wip New: * Next call: Nov 9, same time ==== Meeting 2021/10/11 ==== === Attendees === Jan-Simon, Walt, Scott, Marius ===Topics=== * lava lab status * lava-docker * 10/12 had brief downtime due to power outage - fixed (lab-baylibre) * pyagl: * -next branch / Scott * 9/14 * sllin and qtwayland ftbs fixed * sllin fix can be added to master * qtwayland needs to be sorted by IC group * 10/12 * qtwayland patch disabled in -next , IC Group needs to sort out * sandbox or snapshot for when honister is branched (for reference) * Javier: update on openQA+lava - got POC working * 8/31 * WIP * 9/14 * no update * 10/12 * -> Presentation @ELC and upcoming ALS * Discussion on required changes for CI * build targets * agl-demo-platform(-crosssdk) - alias agl-image-ivi or agl-image-weston * lava tests to change * agl-compositor/screenshooter: minimal change around systemd unit * remove pyagl * remove application-lifecycle * jenkins-job-builder vs gitlab-runner exec ? * next spring 3.5 prserv/hashserv (rw+ro) WIP: * Need to investigate "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * 10/12: vacation New: Next call: Oct 26, same time ==== Meeting 2021/09/14 ==== === Attendees === Jan-Simon, Walt, Scott, Marius, Ishii-san ===Topics=== * lava lab status * lava-docker * 9/14 all up * pyagl: * 9/14 * 1 merged, 2 in review * -next branch / Scott * 8/31 * next branch updated after new syntax change merged, rebased meta-agl-devel/meta-agl-demo as well * agl-demo-platform does not build * sllin not building (likely header diff) * qtwayland ftbs * disable IC qtwayland patch and mention in SAT/IC that fix is needed * 9/14 * sllin and qtwayland ftbs fixed * sllin fix can be added to master * qtwayland needs to be sorted by IC group * Javier: update on openQA+lava - got POC working * 8/31 * WIP * 9/14 * no update WIP: * Marius: testing of IC or non-IVI images * IC , Connectivity, Container/Mesh EG need to drive * interested in testing 'remoting' * 1st: in the next cycle, edi can test the h3+pi4 with the IVI w/ agl-demo-preload + cluster w/ agl-demo-preload * 8/31: Scott will try * 9/14: wip * 2nd: lets find out how we can test this on one board only (h3 and if possible qemu) New: * Need to investigate "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * 9/14: not working, yet. wip. * Discussion on upcoming build targets and ci needs (appfw / flutter) Next call: Oct 12, same time ==== Meeting 2021/08/31 ==== === Attendees === Jan-Simon, Walt, Scott, Marius, Javier ===Topics=== * lava lab status * lava-docker * 8/17 - all up, one rpi4 needs a powered usb hub (rtl-sdr), Corentin will add hub * 8/31 - monitoring pi4 * pyagl: * 8/17 * 3 changes still in gerrit, Edi will look * 8/31 * 2 changes in gerrit, 1 new * -next branch / Scott * 8/31 * next branch updated after new syntax change merged, rebased meta-agl-devel/meta-agl-demo as well * agl-demo-platform does not build * sllin not building (likely header diff) * qtwayland ftbs * disable IC qtwayland patch and mention in SAT/IC that fix is needed * prserv: * 8/31 * thanks to Paul and Scott we can now declare victory. * Usage on -next and after uprev in Q1/22 * docker+ipv6 patch pending on ML * Javier: update on openQA+lava - got POC working * 8/31 * WIP WIP: * Marius: testing of IC or non-IVI images * IC , Connectivity, Container/Mesh EG need to drive * interested in testing 'remoting' * 1st: in the next cycle, edi can test the h3+pi4 with the IVI w/ agl-demo-preload + cluster w/ agl-demo-preload * 8/31: Scott will try * 2nd: lets find out how we can test this on one board only (h3 and if possible qemu) New: * Need to investigate "runqemu audio" - since qemu 4.2 , the setup of audio devices seems to have changed and runqemu uses the old setup. Effect - we can't get audio in and out with runqemu now. * Hints: https://mathiashueber.com/virtual-machine-audio-setup-get-pulse-audio-working/ * Next call: Sept 14, same time ==== Meeting 2021/08/17 ==== === Attendees === Jan-Simon, Walt, Ishii-san, Scott, Javier, Zhou, ===Topics=== * lava lab status * lava-docker * 8/3 - all up, no updates * 8/17 - all up, one rpi4 needs a powered usb hub (rtl-sdr), Corentin will add hub * pyagl: * 8/3 * no update * 8/17 * 3 changes still in gerrit, Edi will look * -next branch / Scott * 8/3: * The syntax change for "overrides" has been changed in master of YP * See: https://docs.yoctoproject.org/migration-guides/migration-3.4.html?highlight=overrides#override-syntax-changes * dunfell compat backport not before 3.1.11 (late september) * TLDR: releases in september might have to be delayed due to that * to keep -next active, we'd have to move our master to dunfell >= 9c2186108e2e8fcf5837c0ffe9dbdeeaea4b7398 for now * 3.1.10 is option for koi and lamprey * prserv: * first issue was asyncio loop in multiple processes * same on client side * shows especially on python3.9 * fix works on python3.9 but still fails on python3.7 * Javier: update on openQA+lava - got POC working WIP: * Marius: testing of IC or non-IVI images * IC , Connectivity, Container/Mesh EG need to drive * interested in testing 'remoting' * 1st: in the next cycle, edi can test the h3+pi4 with the IVI w/ agl-demo-preload + cluster w/ agl-demo-preload * 2nd: lets find out how we can test this on one board only (h3 and if possible qemu) * Next call: Aug 31, same time ==== Meeting 2021/08/03 ==== === Attendees === Jan-Simon, Scott, Zhou, Marius, Qiu ===Topics=== * lava lab status * lava-docker * 8/3 - all up, no updates * pyagl: * 8/3 * no update * -next branch / Scott * 7/06: * todo: re-use yocto-check-layer-wrapper from AB ?! * Scott will check * 7/20: * 2 srcrev updates required for -next to build * Upstream will likely cut their M2 this or next week * Upstream looks into a syntax change ... _append -> :append * 8/3: * The syntax change for "overrides" has been changed in master of YP * See: https://docs.yoctoproject.org/migration-guides/migration-3.4.html?highlight=overrides#override-syntax-changes * dunfell compat backport not before 3.1.11 (late september) * TLDR: releases in september might have to be delayed due to that * to keep -next active, we'd have to move our master to dunfell >= 9c2186108e2e8fcf5837c0ffe9dbdeeaea4b7398 for now * 3.1.10 is option for koi and lamprey * Javier: update on openQA+lava - got POC working WIP: * Marius: testing of IC or non-IVI images * IC , Connectivity, Container/Mesh EG need to drive * interested in testing 'remoting' * 1st: in the next cycle, edi can test the h3+pi4 with the IVI w/ agl-demo-preload + cluster w/ agl-demo-preload * 2nd: lets find out how we can test this on one board only (h3 and if possible qemu) * prserv: * issue on Autobuilder - reproduced issue locally. asyncio code issue, scott instrumenting. New topics: * homescreen-service restart issue Next call: Aug 17, same time ==== Meeting 2021/07/20 ==== === Attendees === Jan-Simon, Zhou Mingying, Scott Murray, Marius Vlad, Virginie Delestree, Corentin Labbe, Kurokawa-san, ===Topics=== * lava lab status * lava-docker * pyagl: * Update 07/06: * small updates * Fujitsu changes in gerrit in review * 07/20 * Edi updated the documentation prior to Lamprey * -next branch / Scott * 7/06: * todo: re-use yocto-check-layer-wrapper from AB ?! * Scott will check * 7/20: * 2 srcrev updates required for -next to build * Upstream will likely cut their M2 this or next week * Upstream looks into a syntax change ... _append -> :append * Javier: update on openQA+lava - got POC working WIP: * fossology.automotivelinux.org: * resetup done and re-upload in progress New topics: * Lamprey and Koi releases this week * Marius: testing of IC or non-IVI images * IC , Connectivity, Container/Mesh EG need to drive * interested in testing 'remoting' * 1st: in the next cycle, edi can test the h3+pi4 with the IVI w/ agl-demo-preload + cluster w/ agl-demo-preload * 2nd: lets find out how we can test this on one board only (h3 and if possible qemu) Next call: Aug 3rd, same time ==== Meeting 2021/07/06 ==== === Attendees === Jan-Simon, Walt, Kurokawa-san, Corentin, Scott, Zhou Minying, ===Topics=== * Corentin: * Update 07/06 * Past week * need to fixup qemu-system-arm's sound * remaining: segfault in agl-service-audiomixer (SPEC-3976) * done * enable upload of kernelci metadata to AGL's instance * need to deploy token in jenkins asap * pyartiproxy * in place * This week * worker (lab) errors out if it looses connection to the master for some time * reproduced in core-lab and lab-baylibre * need to add dependency to curl (agl-netboot) -> jira+gerrit * pyagl: * Update 07/06: * small updates * Fujitsu changes in gerrit in review * -next branch / Scott * 7/06: * todo: re-use yocto-check-layer-wrapper from AB ?! * Scott will check * Javier: update on openQA+lava - got POC working WIP: * fossology.automotivelinux.org: need to change instance size (2vCPU -> 4vCPU) , uprev to fossology 3.10 * instance size changed, no luck. no help/idea from upstream on unbreaking * -> need to resetup New topics: * Next call: Jul 20th, same time ==== Meeting 2021/06/22 ==== === Attendees === Jan-Simon, Walt, Kurokawa-san, Corentin, Virginie, Marius, Scott, Javier, Bernard, ===Topics=== * Corentin: * Update 06/22 * Past week * working on uploads of journals to artifactorial in case of failure * need to fixup qemu-system-arm's sound [ 84.506491] snd_hda_intel 0000:00:02.0: Rejecting highmem page from CMA. [ 84.508426] snd_hda_intel 0000:00:02.0: Rejecting highmem page from CMA. [ 84.511842] snd_hda_intel 0000:00:02.0: Rejecting highmem page from CMA. [ 84.512862] ALSA pcmC0D0p,0:Generic Analog: cannot preallocate for size 65536 [ 84.515193] snd_hda_intel 0000:00:02.0: Rejecting highmem page from CMA. [ 84.532603] snd_hda_intel 0000:00:02.0: Rejecting highmem page from CMA. [ 84.541605] snd_hda_intel 0000:00:02.0: Rejecting highmem page from CMA. [ 84.542165] ALSA pcmC0D0c,0:Generic Analog: cannot preallocate for size 65536 * issue is CONFIG_LPAE in MACHINE=qemuarm, turned config off * remaining: segfault in agl-service-audiomixer (SPEC-3976) * enable upload of kernelci metadata to AGL's instance * can go ahead and merge. * This week * enable upload of kernelci metadata to AGL's instance * can go ahead and merge. * * pyagl: * Update 06/08: * small updates * Fujitsu changes in gerrit in review * -next branch / Scott * 6/08: * todo: re-use yocto-check-layer-wrapper from AB ?! * * gitlab moving ahead * 6/22 not going to happen fast WIP: * fossology.automotivelinux.org: need to change instance size (2vCPU -> 4vCPU) , uprev to fossology 3.10 * older version have issue when batch-uploading (extraction/load/lockups) New topics: * Javier: update on openQA+lava - got POC working Next call: Jul 6th, same time ==== Meeting 2021/06/08 ==== === Attendees === Jan-Simon, Walt, Scott, Corentin, Kurokawa-san ===Topics=== * Corentin: * Update 06/08 * Past week * pyartiproxy: extending now to dump journal or composior.log + report URL to gerrit, can add e.g. zip'ed html output from gcov-wrapper next * update lava-docker with fixes for environment changes (core lab) * KernelCI token (-> Ryan) * check hdmi on h3+kf (failing screenshooter) - done * This week * working on uploads of journals to artifactorial i ncase of failure * need to fixup qemu-system-arm's sound [ 84.506491] snd_hda_intel 0000:00:02.0: Rejecting highmem page from CMA. [ 84.508426] snd_hda_intel 0000:00:02.0: Rejecting highmem page from CMA. [ 84.511842] snd_hda_intel 0000:00:02.0: Rejecting highmem page from CMA. [ 84.512862] ALSA pcmC0D0p,0:Generic Analog: cannot preallocate for size 65536 [ 84.515193] snd_hda_intel 0000:00:02.0: Rejecting highmem page from CMA. [ 84.532603] snd_hda_intel 0000:00:02.0: Rejecting highmem page from CMA. [ 84.541605] snd_hda_intel 0000:00:02.0: Rejecting highmem page from CMA. [ 84.542165] ALSA pcmC0D0c,0:Generic Analog: cannot preallocate for size 65536 * enable upload of kernelci metadata to AGL's instance * pyagl: * Update 06/08: * small updates * -next branch / Scott * 6/08: * todo: re-use yocto-check-layer-wrapper from AB ?! * gitlab moving ahead * 6/08 no news Revisit: * Qui Tingting: gerrit/26226 * need to update to state used in master, e.g. reference pyagl, use pyagl on desktop to remote target * https://docs.automotivelinux.org/en/master/#5_Component_Documentation/7_pyagl/ * https://git.automotivelinux.org/src/pyagl/tree/README.md WIP: * fossology.automotivelinux.org: need to change instance size (2vCPU -> 4vCPU) , uprev to fossology 3.10 * older version have issue when batch-uploading (extraction/load/lockups) New topics: * Next call: Jun 22nd, same time ==== Meeting 2021/05/11 ==== === Attendees === Jan-Simon, Walt, Virginie, Mark, Kurokawa-san, Scott, Thou, Laurence, Guy, Javier ===Topics=== * Corentin: * Update 04/27 * Past week * exporting logs/files from the DUT / test run -> pyartiproxy & artifactorial in place * qemu-system-arm fixes * bbe lite deployed * kernelci.automotivelinux.org metadata fix * This week * pyartiproxy: extending now to dump journal or composior.log + report URL to gerrit, can add e.g. zip'ed html output from gcov-wrapper next * add test runs and evaluate bbe * update lava-docker with fixes for environment changes (core lab) * KernelCI token (-> Ryan) * check hdmi on h3+kf (failing screenshooter) * pyagl: * Update 04/27: * Needs update after pipewire update: * See https://lava.automotivelinux.org/scheduler/job/18137#L4378 * https://jira.automotivelinux.org/browse/SPEC-3898 * 05/11: * edi fixed pyagl after the pipewire update * SPEC-3764 * vmap issue with screenshooter on qemuarm * https://jira.automotivelinux.org/browse/SPEC-3764 * 05/11: * we updated qemu-system-arm, adapted vmap=512M * screenshooter * We need to check the h3+kf (no edid data), HDMI jig * -next branch / Scott * 4/27: * fixes for Autobuilder * atm based on YP 3.3 * yocto-check-layer improvements * https://jira.automotivelinux.org/browse/SPEC-3897 * 5/11: * one build failues due to the meta-pipewire update (lua dependency) * re-use yocto-check-layer-wrapper from AB ?! * gitlab moving ahead * 4/27 no news * 5/11 no news * Marius: https://jira.automotivelinux.org/browse/SPEC-3786 , looking into script (using afb-client-demo) to switch between multiple (web) apps for debugging, can be reused in ci * 4/27: * no update * 5/11: * https://gerrit.automotivelinux.org/gerrit/c/src/agl-shell-activator/+/26330 * Marius noted that https://www.kernel.org/doc/html/v4.12/media/v4l-drivers/vivid.html can help testing video input and also SDR * Qui Tingting: gerrit/26226 * need to update to state used in master, e.g. reference pyagl, use pyagl on desktop to remote target * https://docs.automotivelinux.org/en/master/#5_Component_Documentation/7_pyagl/ * https://git.automotivelinux.org/src/pyagl/tree/README.md New topics: * fossology.automotivelinux.org: need to change instance size (2vCPU -> 4vCPU) , uprev to fossology 3.10 * older version have issue when batch-uploading (extraction/load/lockups) * Next call: May 25th, same time ==== Meeting 2021/04/27 ==== === Attendees === Jan-Simon, Walt, Javier, Virginie, Corentin, Kurokawa-san, Laurence, Mark, Zhou ===Topics=== * Corentin: * Update 04/27 * Past week * Replaced up² board * Checking * ARTIFACTORIAL deployed * This week * Need to fix worker node issues / network * nbd and rootfs timeouts !! * likely need to replace worker (thu) * pyagl: * Update 04/27: * intermittent: https://lava.automotivelinux.org/scheduler/job/17748 (hvac-set) * only indication is a network resetup * WIP, need someone to dig in or disable/skip * https://jira.automotivelinux.org/browse/SPEC-3899 * Needs update after pipewire update: * See https://lava.automotivelinux.org/scheduler/job/18137#L4378 * https://jira.automotivelinux.org/browse/SPEC-3898 * SPEC-3764 * vmap issue with screenshooter on qemuarm * https://jira.automotivelinux.org/browse/SPEC-3764 * -next branch / Scott * 4/27: * fixes for Autobuilder * atm based on YP 3.3 * yocto-check-layer improvements * https://jira.automotivelinux.org/browse/SPEC-3897 * Scott has prepared dunfell/3.1.7 * meta-updater has issues -> devcall * gitlab moving ahead * 4/27 no news * Marius noted: https://gerrit.automotivelinux.org/gerrit/c/apps/agl-service-homescreen/+/26043 * need to check journal wrt service not starting * 4/27: * no update * Marius: https://jira.automotivelinux.org/browse/SPEC-3786 , looking into script (using afb-client-demo) to switch between multiple (web) apps for debugging, can be reused in ci * 4/27: * no update * Marius noted that https://www.kernel.org/doc/html/v4.12/media/v4l-drivers/vivid.html can help testing video input and also SDR New topics: * Qui Tingting: gerrit/26226 * need to update to state used in master, e.g. reference pyagl, use pyagl on desktop to remote target * https://docs.automotivelinux.org/en/master/#5_Component_Documentation/7_pyagl/ * https://git.automotivelinux.org/src/pyagl/tree/README.md Next call: May 11th, same time ==== Meeting 2021/03/30 ==== === Attendees === Jan-Simon, Walt, Corentin, Virginie, zhong liu, Qiu Tingting, Marius, Scott ===Topics=== * Roadmap / Task list review * https://docs.google.com/spreadsheets/d/1WsFPh_-2wAYE8NvKlBGywYixhhJz9EBlbUR02r0UWF8/edit * Corentin: * Update 03/02 * Past week * ARTIFACTORIAL_TOKEN * proxy service + curl push from DUT * try screenshot on all boards * Need to check up² board * power supply or powered usb hub ? * voltage drop under load ? * This week * kernelci build.json fixes (metadata upload) * deploy proxy service in lab * up² final try with new power supply, new board ordered * screenshot on all boards * https://lava.automotivelinux.org/scheduler/alljobs?length=25&search=test-screenshot2#table * need https://gerrit.automotivelinux.org/gerrit/c/AGL/releng-scripts/+/26216 * can enable screenshot * pyagl: * Update 03/30: * all but upsquare seem to work atm * radio fine now * intermittent: https://lava.automotivelinux.org/scheduler/job/17748 (hvac-set) * only indication is a network resetup * -next branch / Scott * 3/02: * set of patches submitted to master * new -next branch populated (repo, meta-agl+meta-agl-demo, few source repos) * target qemu * meta-agl-core in autobuilder: * either DISTRO=poky-agl * or DISTRO=poky plus few settings * AGL_EXTRA_IMAGE_FSTYPES = "" * ... * 3/16: * meta-agl-core in autobuilder: * they'll evaluate poky-agl but might switch to DISTRO=poky. * Form submitted * -next branch usable (building against YP poky master ! for tests and tracking upstream) * Scott rebased -next branch * kernel configuration triggers warnings due to upstream kernel class changes (kernel-yocto.bbclass) * tbd if we disable the extra check or make use of it somehow * working on qemu*, rpi4, already. checking some more platforms * CI needs the rebasing on merge (SPEC-3846) * 3/30 * CI needs the rebasing on merge (SPEC-3846) * gitlab moving ahead, no details, yet. * Marius noted: https://gerrit.automotivelinux.org/gerrit/c/apps/agl-service-homescreen/+/26043 * need to check journal wrt service not starting * 3/30: * no update * Marius: https://jira.automotivelinux.org/browse/SPEC-3786 , looking into script (using afb-client-demo) to switch between multiple (web) apps for debugging, can be reused in ci * 3/30: * no update * Marius noted that https://www.kernel.org/doc/html/v4.12/media/v4l-drivers/vivid.html can help testing video input and also SDR * Li: testing pyagl * Check location of http://old-docs.automotivelinux.org/docs/en/icefish/apis_services/reference/afb-test/0_Installation.html in new doc * Documentation hosted at https://gerrit.automotivelinux.org/gerrit/gitweb?p=AGL/documentation.git;a=tree;h=refs/heads/master;hb=refs/heads/master * 3/30: * Li will look at a patchset New topics: * SPEC-2795 * recommendation is to use pytest * https://git.automotivelinux.org/src/pyagl/tree/ * https://docs.pytest.org/en/stable/ Next call: tentative 27th APR, same time ==== Meeting 2021/03/16 ==== === Attendees === Jan-Simon, Li, Scott Virginie, Corentin, Ed ===Topics=== * Roadmap / Task list review * https://docs.google.com/spreadsheets/d/1WsFPh_-2wAYE8NvKlBGywYixhhJz9EBlbUR02r0UWF8/edit * * Corentin: * Update 03/02 * Past week * try screenshot on all boards * ARTIFACTORIAL_TOKEN * proxy service + curl push from DUT * WIP * LAVA security update * Build metadata for KCi * in review and needs to be enabled * This week * ARTIFACTORIAL_TOKEN * proxy service + curl push from DUT * try screenshot on all boards * Need to check up² board * power supply or powered usb hub ? * voltage drop under load ? * pyagl: * Update 03/02: * issue with usb dongle. * see above * -next branch / Scott * 3/02: * set of patches submitted to master * new -next branch populated (repo, meta-agl+meta-agl-demo, few source repos) * target qemu * meta-agl-core in autobuilder: * either DISTRO=poky-agl * or DISTRO=poky plus few settings * AGL_EXTRA_IMAGE_FSTYPES = "" * ... * 3/16: * meta-agl-core in autobuilder: * they'll evaluate poky-agl but might switch to DISTRO=poky. * Form submitted * -next branch usable (building against YP poky master ! for tests and tracking upstream) * Scott rebased -next branch * kernel configuration triggers warnings due to upstream kernel class changes (kernel-yocto.bbclass) * tbd if we disable the extra check or make use of it somehow * working on qemu*, rpi4, already. checking some more platforms * CI needs the rebasing on merge (SPEC-3846) * gitlab moving ahead, no details, yet. * Marius noted: https://gerrit.automotivelinux.org/gerrit/c/apps/agl-service-homescreen/+/26043 * need to check journal wrt service not starting * Marius: https://jira.automotivelinux.org/browse/SPEC-3786 , looking into script (using afb-client-demo) to switch between multiple (web) apps for debugging, can be reused in ci New: * Marius noted that https://www.kernel.org/doc/html/v4.12/media/v4l-drivers/vivid.html can help testing video input and also SDR * Li: testing pyagl * Check location of http://old-docs.automotivelinux.org/docs/en/icefish/apis_services/reference/afb-test/0_Installation.html in new doc * Documentation hosted at https://gerrit.automotivelinux.org/gerrit/gitweb?p=AGL/documentation.git;a=tree;h=refs/heads/master;hb=refs/heads/master * Li: AGL_APP_REVISION is autorev , for offline builds this means: * need to be on a tag, set BB_NO_NETWORK = "1" , conf/local.conf * if you have an own PREMIRROR, configure it and allow using BB_FETCH_PREMIRROR_ONLY * build target image once, then collect * buildhistory-collect-srcrevs -a * File JIRA as enhancement so we can work on the workflow * -> PREMIRROR -> BB_CREATE_MIRROR_TARBALLS * Li: repo sync: some repo's are slow -> jira On hold: * HTML5 apps, CI integration with ./autobuild/autobuild script Future topics: * 2021 planning: * image/machine matrix today * current agl-demo-platform-crosssdk for most (-nogfx agl-image-ivi-crosssdk) * qemux86-64, qemuarm, qemuaarch64, h3ulcb-nogfx, pi4, bbe * image/machine matrix 2021 * need same as above to keep demo running (qemux86-64, qemuarm, qemuaarch64, h3ulcb-nogfx, pi4) * IC EG image tbd (initially qemux86-64) - container host, container guest (multiconfig) -- * IVI EG image tbd (initially qemux86-64) - minimal image for CI loop -- build ! * telematics-demo / cluster-demo (qemu) - build and boot is fine, full test is problematic * improve testplans for CI, nightly snapshots, next * gitlab-ci use ? Next call: Tuesday Mar 30 th, same timeslot ==== Meeting 2021/03/02 ==== === Attendees === Jan-Simon, Scott, Marius, Corentin, Kurkoawa-san ===Topics=== * Corentin: * Update 03/02 * Past week * change h3ulcb-nogfx to use full graphical image * default to agl-demo-platform-crosssdk * unless branch = jellyfish/halibut * adapt ci-management(ci-platform*) , should be mostly in include-agl-header.sh * and releng-scripts (koi branch, master branch, next branch?) * https://gerrit.automotivelinux.org/gerrit/c/AGL/releng-scripts/+/26083 * ci-management change WIP * radio testing * run manual job executing the pyagl radio tests within baylibre lab to confirm they work * SPEC-3832 * LAVA + KCI updates done * This week * try screenshot on all boards * ARTIFACTORIAL_TOKEN * proxy service + curl push from DUT * pyagl: * 01/05: * radio binding tests: * h3ulcb-nogfx: https://lava.automotivelinux.org/scheduler/job/15266 * pi4: https://lava.automotivelinux.org/scheduler/job/15269 * issue: scan test needs to find other station (aka 1 FM signal not in center signal) * Todo: transmitter send, will arrive later this week, to be installed in lab. * Update 03/02: * issue with usb dongle. * see above * -next branch / Scott * -next branch tracking YP master * need way to uprev necessary layers to track master * -next branches * how do these get bumped * rebase on-top of master (-next uprev is manual, but weekly CI is automated / _floating) * weekly / weekend build * weston/mesa/gfx * qemux86-64 * meta-agl-core in the YP autobuilder * 2/16: * got meta-agl-core building with YP / master - autobuilder use case * some changes can go to master directly * rebase script/procedure wip * next is to build agl-demo-platform (need to see wrt braches involved) * 3/02: * set of patches submitted to master * new -next branch populated (repo, meta-agl+meta-agl-demo, few source repos) * target qemu * meta-agl-core in autobuilder: * either DISTRO=poky-agl * or DISTRO=poky plus few settings * AGL_EXTRA_IMAGE_FSTYPES = "" * ... * gitlab moving ahead, no details, yet. * Marius noted: https://gerrit.automotivelinux.org/gerrit/c/apps/agl-service-homescreen/+/26043 * need to check journal wrt service not starting * Marius: https://jira.automotivelinux.org/browse/SPEC-3786 , looking into script (using afb-client-demo) to switch between multiple (web) apps for debugging, can be reused in ci New: * Marius noted that https://www.kernel.org/doc/html/v4.12/media/v4l-drivers/vivid.html can help testing video input and also SDR * Li: testing pyagl * Check location of http://old-docs.automotivelinux.org/docs/en/icefish/apis_services/reference/afb-test/0_Installation.html in new doc * Documentation hosted at https://gerrit.automotivelinux.org/gerrit/gitweb?p=AGL/documentation.git;a=tree;h=refs/heads/master;hb=refs/heads/master On hold: * HTML5 apps, CI integration with ./autobuild/autobuild script Future topics: * 2021 planning: * image/machine matrix today * current agl-demo-platform-crosssdk for most (-nogfx agl-image-ivi-crosssdk) * qemux86-64, qemuarm, qemuaarch64, h3ulcb-nogfx, pi4, bbe * image/machine matrix 2021 * need same as above to keep demo running (qemux86-64, qemuarm, qemuaarch64, h3ulcb-nogfx, pi4) * IC EG image tbd (initially qemux86-64) - container host, container guest (multiconfig) -- * IVI EG image tbd (initially qemux86-64) - minimal image for CI loop -- build ! * telematics-demo / cluster-demo (qemu) - build and boot is fine, full test is problematic * improve testplans for CI, nightly snapshots, next * gitlab-ci use ? Next call: Tuesday Mar 30 th, same timeslot ==== Meeting 2021/02/16 ==== === Attendees === Jan-Simon Möller, Walt, Scott, Kurokawa-san, Virginie, Corentin, Marius, Anmol ===Topics=== * Corentin: * Update 02/16 * Past week * Screenshot script * need jigs on all boards to work - DONE * ARTIFACTORIAL_TOKEN - no news upstream, patch in local tree/deployed * lava update preparations (replacing zmq with tokens) * update easy and automatic * SPEC- * This week * change h3ulcb-nogfx to use full graphical image * default to agl-demo-platform-crosssdk * unless branch = jellyfish/halibut * adapt ci-management(ci-platform*) , should be mostly in include-agl-header.sh * and releng-scripts (koi branch, master branch, next branch?) * radio testing * run manual job executing the pyagl radio tests within baylibre lab to confirm they work * try screenshot on all boards * pyagl: * 01/05: * radio binding tests: * h3ulcb-nogfx: https://lava.automotivelinux.org/scheduler/job/15266 * pi4: https://lava.automotivelinux.org/scheduler/job/15269 * issue: scan test needs to find other station (aka 1 FM signal not in center signal) * Todo: transmitter send, will arrive later this week, to be installed in lab. * Update 02/16: * radio transmitter deployed, corentin will do tests * -next branch / Scott * -next branch tracking YP master * need way to uprev necessary layers to track master * -next branches * how do these get bumped * rebase on-top of master (-next uprev is manual, but weekly CI is automated / _floating) * weekly / weekend build * weston/mesa/gfx * qemux86-64 * meta-agl-core in the YP autobuilder * 2/16: * got meta-agl-core building with YP / master - autobuilder use case * some changes can go to master directly * rebase script/procedure wip * next is to build agl-demo-platform (need to see wrt braches involved) New: * CI infra issues resolved now. JVM on build nodes crashed. Buildnodes are now 18.04. * gitlab still WIP * agl-demo-platform-html5: https://jira.automotivelinux.org/browse/SPEC-3805 , need recheck on 18.04 * Marius noted: https://gerrit.automotivelinux.org/gerrit/c/apps/agl-service-homescreen/+/26043 * need to check RID being '' (continue vs return) * need to check journal wrt service not starting * Marius: https://jira.automotivelinux.org/browse/SPEC-3801 , main CPU hog is in QT * Marius: https://jira.automotivelinux.org/browse/SPEC-3786 , looking into script (using afb-client-demo) to switch between multiple apps for debugging, can be reused in ci On hold: * gitlab mirror * HTML5 apps, CI integration with ./autobuild/autobuild script Future topics: * 2021 planning: * image/machine matrix today * current agl-demo-platform-crosssdk for most (-nogfx agl-image-ivi-crosssdk) * qemux86-64, qemuarm, qemuaarch64, h3ulcb-nogfx, pi4, bbe * image/machine matrix 2021 * need same as above to keep demo running (qemux86-64, qemuarm, qemuaarch64, h3ulcb-nogfx, pi4) * IC EG image tbd (initially qemux86-64) - container host, container guest (multiconfig) -- * IVI EG image tbd (initially qemux86-64) - minimal image for CI loop -- build ! * telematics-demo / cluster-demo (qemu) - build and boot is fine, full test is problematic * improve testplans for CI, nightly snapshots, next * gitlab-ci use ? Next call: Tuesday Mar 2nd, same timeslot ==== Meeting 2021/02/01 ==== === Attendees === Jan-Simon Möller, Virginie, Scott, Corentin, Paul, Stephane, Marius ===Topics=== * Corentin: * Update 02/16 * Past week * Screenshot script * need jigs on all boards to work - DONE * ARTIFACTORIAL_TOKEN - no news upstream, patch in local tree/deployed * lava update preparations (replacing zmq with tokens) * update easy and automatic * SPEC- * This week * change h3ulcb-nogfx to use full graphical image * default to agl-demo-platform-crosssdk * unless branch = jellyfish/halibut * adapt ci-management(ci-platform*) , should be mostly in include-agl-header.sh * and releng-scripts (koi branch, master branch, next branch?) * radio testing * run manual job executing the pyagl radio tests within baylibre lab to confirm they work * pyagl: * 01/05: * radio binding tests: * h3ulcb-nogfx: https://lava.automotivelinux.org/scheduler/job/15266 * pi4: https://lava.automotivelinux.org/scheduler/job/15269 * issue: scan test needs to find other station (aka 1 FM signal not in center signal) * Todo: transmitter send, will arrive later this week, to be installed in lab. * Update 02/16: * radio transmitter deployed, corentin will do tests * -next branch * -next branch tracking YP master * need way to uprev necessary layers to track master * -next branches * how do these get bumped * rebase on-top of master (-next uprev is manual, but weekly CI is automated / _floating) * weekly / weekend build * weston/mesa/gfx * qemux86-64 * meta-agl-core in the YP autobuilder New: * - On hold: * gitlab mirror * HTML5 apps, CI integration with ./autobuild/autobuild script Future topics: * 2021 planning: * image/machine matrix today * current agl-demo-platform-crosssdk for most (-nogfx agl-image-ivi-crosssdk) * qemux86-64, qemuarm, qemuaarch64, h3ulcb-nogfx, pi4, bbe * image/machine matrix 2021 * need same as above to keep demo running (qemux86-64, qemuarm, qemuaarch64, h3ulcb-nogfx, pi4) * IC EG image tbd (initially qemux86-64) - container host, container guest (multiconfig) -- * IVI EG image tbd (initially qemux86-64) - minimal image for CI loop -- build ! * telematics-demo / cluster-demo (qemu) - build and boot is fine, full test is problematic * improve testplans for CI, nightly snapshots, next * gitlab-ci use ? Next call: Tuesday Feb 16th, same timeslot ==== Meeting 2021/01/19 ==== === Attendees === Jan-Simon Möller, Marius, Virginie, Li, Corentin ===Topics=== * Corentin: * Update 01/19 * Past week * kernelci update * AGL refhw / qemux86-64 jobs on up² * This week * kernelci update * backup/restore patch pending for merge * update procedure for LFIT * AGL refhw / qemux86-64 jobs on up² * working on submitting 2 jobs for each board (h3+kf/agl-refhw and qemux86-64/up²) * Screenshot script * qemu works * h3ulcb-nogfx: exempt from screenshot (investigating pixman+drm or fbdev ?) * for JS: agl-compositor --use-pixman * to be enabled for qemu*, pi4, up² * not enable for *-nogfx for now * pyagl: * 01/05: * radio binding tests: * h3ulcb-nogfx: https://lava.automotivelinux.org/scheduler/job/15266 * pi4: https://lava.automotivelinux.org/scheduler/job/15269 * issue: scan test needs to find other station (aka 1 FM signal not in center signal) * a) Corentin: unplug RTL-SDR usb dongle from pi4 in lab * b) Jan-Simon: send 1 FM transmitter to Corentin * c) look at h3ulcb-nogfx build to see why service is not up * d) verify pi4 locally first to rule out issue in SDR * agl-compositor / screenshotter * 10/13: WIP * screenshotter code in master * recipe changes and script next * 01/19: * WIP, script under test * Script has positive result * We can enable it (2nd last action before journal dump) * -next branch tracking YP master * to be discussed next time. On hold: * gitlab mirror * 12/08 * next: need to setup project specific runner * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * 09/01 * html5 on hold New: * 2021 planning: * image/machine matrix today * current agl-demo-platform-crosssdk for most (-nogfx agl-image-ivi-crosssdk) * qemux86-64, qemuarm, qemuaarch64, h3ulcb-nogfx, pi4, bbe * image/machine matrix 2021 * need same as above to keep demo running (qemux86-64, qemuarm, qemuaarch64, h3ulcb-nogfx, pi4) * IC EG image tbd (initially qemux86-64) - container host, container guest (multiconfig) -- * IVI EG image tbd (initially qemux86-64) - minimal image for CI loop -- build ! * telematics-demo / cluster-demo (qemu) - build and boot is fine, full test is problematic * -next branch tracking YP master * need way to uprev necessary layers to track master * -next branches * how do these get bumped * rebase on-top of master (-next uprev is manual, but weekly CI is automated / _floating) * weekly / weekend build * weston/mesa/gfx * qemux86-64 * improve testplans for CI, nightly snapshots, next * gitlab-ci use ? * Marius: difference between bitbake -c populate_sdk agl-demo-platform (matching crosssdk for specified image) and bitbake agl-demo-platform-crosssdk (image and sdk as superset of above and workaround for meta-qt5) Next call: Tuesday Feb 2nd, same timeslot ==== Meeting 2020/12/08 ==== === Attendees === Jan-Simon Möller, Scott, Corentin, Li, Virginie ===Topics=== * Corentin: * Update 01/05 * Past week * This week * kernelci update * database restore work * update procedure * 01/05: WIP * AGL refhw * Healthcheck job * 01/05: h3ulcb-nogfx can drive refhw lava tests * qemux86-64 jobs on up² * releng and lava bits in preparation * limit to master and jellyfish (patches for icefish in prep) * 01/05: ci-management repo needs to enable pushing jobs * pyagl: * todo: add peripheral hardware * Corentin has usb RTLSDR dongles * pi4 * h3+kf * up² * refhw * 12/08: * agl-service-hvac added to pyagl * Scott fixing agl-service-radio in gerrit * master and with additional changes in jellyfish * 01/05: * radio binding tests: * h3ulcb-nogfx: https://lava.automotivelinux.org/scheduler/job/15266 * pi4: https://lava.automotivelinux.org/scheduler/job/15269 * issue: scan test needs to find other station (aka 1 FM signal not in center signal) * a) Corentin: unplug RTL-SDR usb dongle from pi4 in lab * b) Jan-Simon: send 1 FM transmitter to Corentin * c) look at h3ulcb-nogfx build to see why service is not up * d) verify pi4 locally first to rule out issue in SDR * agl-compositor / screenshotter * 10/13: WIP * screenshotter code in master * recipe changes and script next * 01/05: * need to ask Marius * gitlab mirror * 12/08 * next: need to setup project specific runner On hold: * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * 09/01 * html5 on hold New: * 2021 planning: * image/machine matrix today * current agl-demo-platform-crosssdk for most (-nogfx agl-image-ivi-crosssdk) * qemux86-64, qemuarm, qemuaarch64, h3ulcb-nogfx, pi4, bbe * image/machine matrix 2021 * need same as above to keep demo running (qemux86-64, qemuarm, qemuaarch64, h3ulcb-nogfx, pi4) * IC EG image tbd (initially qemux86-64) - container host, container guest (multiconfig) -- * IVI EG image tbd (initially qemux86-64) - minimal image for CI loop -- build ! * telematics-demo / cluster-demo (qemu) - build and boot is fine, full test is problematic * -next branch tracking YP master * need way to uprev necessary layers to track master * -next branches * how do these get bumped * rebase on-top of master (-next uprev is manual, but weekly CI is automated / _floating) * weekly / weekend build * weston/mesa/gfx * qemux86-64 * improve testplans for CI, nightly snapshots, next * gitlab-ci use ? * Next call: Tuesday Jan 19th, same timeslot ==== Meeting 2020/11/10 ==== === Attendees === Jan-Simon Möller, Scott Murray, Corentin Labbe ===Topics=== * Corentin: * Update 11/10 * Past week * re-enable upsquare * board reboots during bootup, might be kernel CONFIG_ (missing?) * agl-service-geoclue, agl-service-weather fixes jellyfish * more token injection / upstream * This week * kernelci update * AGL refhw for lab-baylibre * jellyfish: pytest can failures on h3 in CI * backport of can move * should fix jellyfish in CI * re-enable upsquare * Details in SPEC-2962 * plan: * wait for YP point release uprev in AGL * eval kernel update * pyagl: * signal-composer disabled * rds code * todo: add peripheral hardware * Corentin has usb RTLSDR dongles * pi4 * h3+kf * up² * refhw * agl-compositor / screenshotter * 10/13: WIP * screenshotter code in master * recipe changes and script next * * gitlab mirror * 11/10 * todo: autocloser actions * Test wgts: * disabled in CI due to test wgts killing homescreen. using pyagl. * details in SPEC-3608 * see agl-service-homescreen json refcount bug (SPEC-3694) On hold: * SW360: * Update 09/15 * sw360.zip: liferay start does not complete * needs some work to get this up * upstream: recent upstream did not work * https://github.com/sw360/sw360chores * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * 09/01 * html5 on hold New: Next call: Tuesday Nov 24th ==== Meeting 2020/09/29 ==== === Attendees === Jan-Simon Möller, Walt, Corentin, Scott, Marius, Matthew ===Topics=== * Corentin: * Update 10/13 * Past week * fixing qemu vga, sound * WIP: env var injection / token injection for artifact upload * upstream https://git.lavasoftware.org/lava/lava/-/merge_requests/1342 * This week * more token injection / upstream * qemu network if updates * pi4 u-boot fdt additions * pi need veth when in nbd boot * kernelci update * gcov (for -coverage wgt run on target) * 10/13: live in CI, follow-up with report to gerrit ? * pyagl: * 2 more CAN tests missing * signal-composer disabled * rds code * in CI working out infra issues * e.g. all boards need network visible to network service * todo: add most hardware to h3 in corelab * agl-compositor / screenshotter * 10/13: WIP * screenshotter code in master * recipe changes and script next * gitlab mirror * 09/15 WIP to mirror out to gitlab.com/automotivegradelinux * 10/13 * todo: autocloser actions * AGL license scanning (SPDX/fossology) * 09/29 * 10.0.0 uploaded * Test wgts: * disabled in CI due to test wgts killing homescreen. using pyagl. * details in SPEC-3608 On hold: * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Update 08/18 * next: * need to update kernelci instance * SW360: * Update 09/15 * sw360.zip: liferay start does not complete * needs some work to get this up * upstream: recent upstream did not work * https://github.com/sw360/sw360chores * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * 09/01 * html5 on hold New: * Workshop post discussion * Scott: h3ulcb-nogfx w/ or w/o kingfisher: no kingfisher layer in there Next call: Tuesday Nov 10th ==== Meeting 2020/09/29 ==== === Attendees === Jan-Simon Möller, Corentin ===Topics=== * Corentin: * Update 09/15 * Past week * -coverage * pytest / pyagl integration * This week * fix qemu wrt "-device VGA,vgamem_mb=64,edid=on" * check also: -device qemu-xhci -device usb-tablet -device usb-kbd -object rng-random,filename=/dev/urandom,id=rng0 -device virtio-rng-pci,rng=rng0 * qemu-system-aarch64 -drive id=disk0,file=agl-demo-platform-crosssdk.ext4,if=none,format=raw -device virtio-blk-device,drive=disk0 __**-show-cursor -device VGA,vgamem_mb=64,edid=on -device qemu-xhci -device usb-tablet -device usb-kbd**__ -object rng-random,filename=/dev/urandom,id=rng0 -device virtio-rng-pci,rng=rng0 -vnc :0 **-soundhw hda** **-machine virt -cpu cortex-a57** -m 2048 -serial mon:stdio -serial null -kernel /home/dl9pf/yocto/build-qemuarm64/tmp/deploy/images/qemuarm64/Image--5.4.57+git0+4aeda12f7f_d3100abbe1-r0-qemuarm64-20200928084041.bin -append 'root=/dev/vda rw console=ttyS0 mem=2048M ip=dhcp console=ttyAMA0 quiet ' * env.yaml / env.dut.yaml in lava-docker ? * sound on (env QEMU_AUDIO_DRV=none) in env.yaml * env.dut.yaml (deploy with lava-docker) - env vars for artifactorial * kernelci update * gcov (for -coverage wgt run on target) * issue: gcov files from the build need to be present on target * gcda and gcno need to be in a build-time directory structure * build-time directory structure is embedded * might need gcovr (or lcov) , jq ? * GCOV_PREFIX, GCOV_PREFIX_STRIP to be set in env variable * GCOV_PREFIX=/tmp/gcov// * option 1: -fprofile-prefix-path ? (but gcc-10 only or patch) * option 2: namespace + overlayfs to deploy necessary folder structure * or kcov (but bad with shared libraries https://github.com/SimonKagstrom/kcov ) * deploy: * /usr/src/debug/ (default debugsource location in YP) * /usr/src/coverage/ * run -coverage wgt * mkdir -p /tmp/gcov/ * cp -ar /usr/src/coverage//* /tmp/gcov/ * chsmack -r -a "System::shared" /tmp/gcov * stop -coverage wgt (to write out files) * find .gcda files and run gcov in source directory * report coverage % through lava result * go forward with deploy part, test -fprofile-prefix-path patch (app-templates) * gitlab mirror * 09/15 WIP to mirror out to gitlab.com/automotivegradelinux * 09/29 * mirror in active sync * todo: autocloser actions * AGL license scanning (SPDX/fossology) * 09/15 * 9.0.3 uploaded, ready for clearing * 09/29 * 10.0.0 uploaded * Test wgts: * Need help going through wgts and making sure * afm-util install/uninstall * afm-util stop/start * basically: https://git.automotivelinux.org/src/qa-testdefinitions/tree/common/scripts/application-lifecycle.sh * qemux86-64 * wgt tests 'passlist': * - cloud-proxy service * - sometimes network service * - bluetooth (start race) On hold: * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Update 08/18 * next: * need to update kernelci instance * SW360: * Update 09/01: * Temporary location: http://95.217.112.185:8080/ * Update 09/15 * sw360.zip: liferay start does not complete * needs some work to get this up * upstream: recent upstream did not work * https://github.com/sw360/sw360chores * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * 09/01 * html5 on hold New: * Next call: Tuesday Oct 13th - same timeslot. ==== Meeting 2020/09/15 ==== === Attendees === Jan-Simon Möller, Corentin, Scott, Li, Brian, Matthew, Takada-san, Zhou ===Topics=== * Fuego / Lava support * Update 09/15 * https://git.lavasoftware.org/lava/lava/-/merge_requests/330 merged ! * Corentin: * Update 09/15 * Past week * https://git.lavasoftware.org/lava/lava/-/merge_requests/330 * extend for -coverage * This week * -coverage * pytest / pyagl integration * kernelci update * gcov (for -coverage wgt run on target) * issue: gcov files from the build need to be present on target * gcda and gcno need to be in a build-time directory structure * build-time directory structure is embedded * might need gcovr (or lcov) , jq ? * GCOV_PREFIX, GCOV_PREFIX_STRIP to be set in env variable * GCOV_PREFIX=/tmp/gcov// * option 1: -fprofile-prefix-path ? (but gcc-10 only or patch) * option 2: namespace + overlayfs to deploy necessary folder structure * or kcov (but bad with shared libraries https://github.com/SimonKagstrom/kcov ) * deploy: * /usr/src/debug/ (default debugsource location in YP) * /usr/src/coverage/ * run -coverage wgt * mkdir -p /tmp/gcov/ * cp -ar /usr/src/coverage//* /tmp/gcov/ * chsmack -r -a "System::shared" /tmp/gcov * stop -coverage wgt (to write out files) * find .gcda files and run gcov in source directory * report coverage % through lava result * go forward with deploy part, test -fprofile-prefix-path patch (app-templates) * gitlab mirror * 09/15 WIP to mirror out to gitlab.com/automotivegradelinux * python-driven tests for appfw (pyagl) * 09/01 plan is to have a first patchset in gerrit this week * 09/15 merged. Work on CI integration * AGL license scanning (SPDX/fossology) * 09/15 * 9.0.3 uploaded, ready for clearing * Test wgts: * Need help going through wgts and making sure * afm-util install/uninstall * afm-util stop/start * basically: https://git.automotivelinux.org/src/qa-testdefinitions/tree/common/scripts/application-lifecycle.sh * qemux86-64 * wgt tests 'passlist': * - radio service cannot work in qemu * do a null driver done * 9/15 working on tests * - cloud-proxy service * - sometimes network service * - bluetooth (start race) On hold: * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Update 08/18 * next: * need to update kernelci instance * SW360: * Update 09/01: * Temporary location: http://95.217.112.185:8080/ * Update 09/15 * sw360.zip: liferay start does not complete * needs some work to get this up * upstream: recent upstream did not work * https://github.com/sw360/sw360chores * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * 09/01 * html5 on hold New: * Corentin: Jan-Simon will look for SDR dongle for UP2 * Li: Next call: Tuesday Sep 29th - same timeslot. ==== Meeting 2020/09/01 ==== === Attendees === Jan-Simon Möller, Walt Miner, Li Xiaoming, Corentin Labbe, Scott Murray, Matthew Commons ===Topics=== * Fuego / Lava support * Update 09/01 * Tested with latest master, hope to get it in upstream * https://git.lavasoftware.org/lava/lava/-/merge_requests/330 * Corentin: * Update 09/01 * Past week * extend for -coverage * https://git.lavasoftware.org/lava/lava/-/merge_requests/330 * lava uprev * This week * -coverage * kernelci update * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * 08/18: * Need to follow-up and ship the source tree with the coverage wgt. * need to run gcov on the target when the trace is finished * thus ship source tree with gcov annotation files * 09/01 * html5 on hold * python-driven tests for appfw * 09/01 plan is to have a first patchset in gerrit this week * AGL license scanning (SPDX/fossology) * 07.07: * 9.0.2 analysis did run * open for license review * 08/28: * 9.0.2 done in fossology * 9.0.3 will be uploaded when released later this week * should we try the reuse decider ? * Li: yes for AGL packages * 09/01: * 9.0.3 uploading to fossology * will send email when ready for license clearing. * SW360: * Will send Dockerfile for sw360 container * Update 06/09: * https://wiki.automotivelinux.org/_media/eg-ciat/sw360.zip new file * AI JS: Check with Ryan how to deploy * Update 07/07 * Host is up, port not exposed, yet. * Update 07/21 * No update * Update 08/18: * Host updates pending / disk space. * Update 09/01: * Temporary location: http://95.217.112.185:8080/ * will get domain next * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click * Update 07/07 * blocked by python 3 update * Update 21/07 * unblocked wrt python3 update * Update 08/18 * next: * need to update kernelci instance * Test wgts: * Need help going through wgts and making sure * afm-util install/uninstall * afm-util stop/start * basically: https://git.automotivelinux.org/src/qa-testdefinitions/tree/common/scripts/application-lifecycle.sh * qemux86-64 * Jan-Simon will contact Li wrt update SPEC-2795 * wgt tests 'passlist': * - radio service cannot work in qemu * option1: make it pass * option2: do a null driver * - cloud-proxy service * - sometimes network service * - bluetooth (start race) New: * Pawel: * Brought up Board@Desk during AMM. Investigating libguestfs issue, will demo later. * Corentin: Jan-Simon will look for SDR dongle for UP2 Next call: Tuesday Sep 15th - same timeslot. ==== Meeting 2020/08/18 ==== === Attendees === Jan-Simon Möller, Mathew Commons, Corentin Labbe, Seigo Takada, Scott Murray, Marius Vlad ===Topics=== * Fuego / Lava support * Update 07/07 * Patch updated in https://git.lavasoftware.org/lava/lava/-/merge_requests/330 * Remi OOO till Jul 30th * Update 08/18 * No update * Corentin: * Update 08/18 * Past week * building multiple wgt packages in CI * updated application lifecycle to run 'release', 'test', 'debug' in CI * This week * extend for -coverage * https://git.lavasoftware.org/lava/lava/-/merge_requests/330 jira) * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * 07/21: * Follow-up: ship source tree with coverage wgt to allow parsing on the target (gcov) * does the debug/coverage prefix strip option end-up in the SDK ??? * 08/18: * Need to follow-up and ship the source tree with the coverage wgt. * need to run gcov on the target when the trace is finished * thus ship source tree with gcov annotation files * AGL license scanning (SPDX/fossology) * 07.07: * 9.0.2 analysis did run * open for license review * 08/28: * 9.0.2 done in fossology * 9.0.3 will be uploaded when released later this week * should we try the reuse decider ? * Li: yes for AGL packages * SW360: * Will send Dockerfile for sw360 container * Update 06/09: * https://wiki.automotivelinux.org/_media/eg-ciat/sw360.zip new file * AI JS: Check with Ryan how to deploy * Update 07/07 * Host is up, port not exposed, yet. * Update 07/21 * No update * Update 08/18: * Host updates pending / disk space. * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click * Update 07/07 * blocked by python 3 update * Update 21/07 * unblocked wrt python3 update * Update 08/18 * next: * need to update kernelci instance * plan is to do this during the next planned downtime in ~2weeks * prep branch or tag until then * Li: how to test binding manually (post/token) * afaik afm-util show id will tell port * if agl-devel is on, then check journalctl * 05/12: no update * 05/26: Li did some more tests, issue/jira ? * Jan 24 14:49:47 raspberrypi3 afm-system-daemon[292]: ERROR: can't start system unit afm-appli-myapp--0.1--main@.service for uid 1001 [/w/workspace/release-jjb-halibut-snapshot/MACHINE/raspberrypi3/label/agl-test-slave/repoclone/output/tmp/work/armv7vet2hf-neon-vfpv4-agl-linux-gnueabi/af-main/halibut+gitAUTOINC+1289070816-r0/git/src/afm-urun.c:252] * 06/09: * no update * 06/23: * https://lava.automotivelinux.org/scheduler/job/10818#results_1029988 * 08/18: * We're seeing a lot of issues when we run the in CI now. * Test wgts need inspection on issues: * FAIL: Failed to start application agl-service-geofence-test * ERROR: failed Resource temporarily unavailable * Update 05/26: WIP * Update 06/09: * https://jira.automotivelinux.org/browse/SPEC-3365 * No further updates * 06/23: * https://jira.automotivelinux.org/browse/SPEC-3365 * no updates * 07/07 * no update * 07/21 * https://gerrit.automotivelinux.org/gerrit/c/src/app-framework-main/+/25037 * Fixes: afm-util run (no 2nd restart) - and test wgt issues * 08/18 * partially fixed, still test wgts need a check New: * Pawel: * Brought up Board@Desk during AMM. Investigating libguestfs issue, will demo later. * Next call: Tuesday Sep 1st - same timeslot. ==== Meeting 2020/07/21 ==== === Attendees === Jan-Simon Möller, Scott Murray, Pawel Wieczorek, Mathew Commons, Li Xiaoming, ===Topics=== * Fuego / Lava support * Update 12/10: * Reworking https://git.lavasoftware.org/lava/lava/merge_requests/330 * waiting for upstream review * Update 07/07 * Patch updated in https://git.lavasoftware.org/lava/lava/-/merge_requests/330 * Remi OOO till Jul 30th * Corentin: * Update 07/07 * Past week * merged python3 support * check app-test (afm-util status) * This week * extend ci-management to run "autobuild package-all" (insert jira) * extent ci-management (application jobs) to upload the multiple wgt files being built now (insert jira) * Application tests , build multiple wgts * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * 07/21: * Follow-up: ship source tree with coverage wgt to allow parsing on the target (gcov) * does the debug/coverage prefix strip option end-up in the SDK ??? * AGL license scanning (SPDX/fossology) * 07.07: * 9.0.2 analysis did run * open for license review * SW360: * Will send Dockerfile for sw360 container * Update 06/09: * https://wiki.automotivelinux.org/_media/eg-ciat/sw360.zip new file * AI JS: Check with Ryan how to deploy * Update 07/07 * Host is up, port not exposed, yet. * Update 07/21 * No update * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click * Update 07/07 * blocked by python 3 update * Update 21/07 * unblocked wrt python3 update * Li: how to test binding manually (post/token) * afaik afm-util show id will tell port * if agl-devel is on, then check journalctl * 05/12: no update * 05/26: Li did some more tests, issue/jira ? * Jan 24 14:49:47 raspberrypi3 afm-system-daemon[292]: ERROR: can't start system unit afm-appli-myapp--0.1--main@.service for uid 1001 [/w/workspace/release-jjb-halibut-snapshot/MACHINE/raspberrypi3/label/agl-test-slave/repoclone/output/tmp/work/armv7vet2hf-neon-vfpv4-agl-linux-gnueabi/af-main/halibut+gitAUTOINC+1289070816-r0/git/src/afm-urun.c:252] * 06/09: * no update * 06/23: * https://lava.automotivelinux.org/scheduler/job/10818#results_1029988 * Test wgts need inspection on issues: * FAIL: Failed to start application agl-service-geofence-test * ERROR: failed Resource temporarily unavailable * Update 05/26: WIP * Update 06/09: * https://jira.automotivelinux.org/browse/SPEC-3365 * No further updates * 06/23: * https://jira.automotivelinux.org/browse/SPEC-3365 * no updates * 07/07 * no update * 07/21 * https://gerrit.automotivelinux.org/gerrit/c/src/app-framework-main/+/25037 * Fixes: afm-util run (no 2nd restart) - and test wgt issues New: * Li: * https://jira.automotivelinux.org/browse/SPEC-3508 * Brian: * Build caching demo * Pawel: * Brought up Board@Desk during AMM. Investigating libguestfs issue, will demo later. Next call: Tuesday Aug 18th - same timeslot. ==== Meeting 2020/07/07 ==== === Attendees === Jan-Simon, Corentin, Li, Scott ===Topics=== * infra: please try responsiveness of * https://confluence.automotivelinux.org/display/MAIN/AGL+Long+Term+Support * vs. * https://test-confluence.automotivelinux.org/display/TEST/AGL+Long+Term+Support * Fujitsu test report update (Fujitsu) * 06/23 * Pi4 no audio ... let's check driver module/kernel config and wireplumber-config device name * we do build CONFIG_SND_BCM2835=m * there is no wireplumber-config for the pi4, so the default might end-up only on hdmi (best guess) * https://git.automotivelinux.org/AGL/meta-agl-devel/tree/meta-pipewire/recipes-multimedia/wireplumber/wireplumber-board-config-agl?h=master * 07/07 * RC1: no test due to budget * upsquare: use qemux86-64 wic * pi: SPEC-3455 for UI (SPEC-3348) * Fuego / Lava support * Update 12/10: * Reworking https://git.lavasoftware.org/lava/lava/merge_requests/330 * waiting for upstream review * Update 05/26 * ported and staged for 2020.05 * Update 06/09: * Corentin will ping Remi * Update 06/23 * No update from Corentin * Update 07/07 * Patch updated in https://git.lavasoftware.org/lava/lava/-/merge_requests/330 * Remi OOO till Jul 30th * Corentin: * Update 07/07 * Past week * Fixed packer images, * This week * merge python3 patches * check app-test (afm-util status) * extend ci-management to run "autobuild package-all" (insert jira) * extent ci-management (application jobs) to upload the multiple wgt files being built now (insert jira) * Application tests , build multiple wgts * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * Update 03/03: * postponed after test wgt * 04/28: * Scott will post his first round of aglwgt/autobuild fixes after next is merged. * Plan is to review in gerrit this week * 05/12: * Scott has his patches for review in gerrit * https://gerrit.automotivelinux.org/gerrit/q/topic:%2522smurray/aglwgt%2522 * First patchset out of a series. * 05/26: * Next set of patches in preparation * 06/23: * aglwgt bbclass rewrite * waiting for marius branch to land first * 7/7: * declare virtory for now * Follow-up: ship source tree with coverage wgt to allow parsing on the target (gcov) * does the debug/coverage prefix strip option end-up in the SDK ??? * AGL license scanning (SPDX/fossology) * 05/12: * License clearing done. * Next step? * 05/26 * Next ? * extend clearing work ? * more boards * or more packages * discussion about completing one target. Li will discuss further. * fossology has a 'reuse' decider that can built on existing decisions from previous runs. * 06/09: * 9.0.2 being uploaded to the fossology server * no analysis run, yet. still making sure all artifacts are fully uploaded. * found issue Dockerfile should install unzip (so we we don't use the php built-in), verifying that. Will have Ryan update our container. * 06/23: * unzip issue still a WIP item as: * during 'ununpack', we cannot abort the agent. But also cannot reupload the same file * the agents can get stuck -> unstable operation * 07.07: * 9.0.2 analysis did run * open for license review * SW360: * Will send Dockerfile for sw360 container * Update 03/03: * eta in 2 weeks * Update 03/31: * tests of sw360 in progress, working on 2 issues atm * Update 04/13: * sw360 Dockerfile and user guide:{{:eg-ciat:sw360.7z|here}}. * Todo Jan-Simon: setup sw360 with ryan * Setup server on same aws host but increase cpu/mem * Go from there. * Update 04/28: * Afraid we cannot use the Dockerfile as-is in AGL infra * Items to work on: * a) no proxy settings, no proxy user, no proxy pass * TLDR: if we deploy this in AWS, there is no Proxy at all. * b) Dockerfile uses downloaded Oracle jdk: can't we use openjdk installed in the distro installed by apt-get? * c) all data/databases need to be on persistent shared volume(s) * d) all setup needs to be automatic (in Dockerfile) * no manual setup steps - only docker build & docker run * e) ? would it make sense to split using docker-compose ? * Who can work on this ? @Kusakabe-san * Update 05/12: * work started on Dockerfile * Update 05/26: * a), b), c) completed * d) wip * let's exchange email * Update 06/09: * https://wiki.automotivelinux.org/_media/eg-ciat/sw360.zip new file * AI JS: Check with Ryan how to deploy * Update 06/23: * Still working through it with * Update 07/07 * Host is up, port not exposed, yet. * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click * Update 06/09 * based on KernelCI instanced update * Update 07/07 * blocked by python 3 update * Li: how to test binding manually (post/token) * afaik afm-util show id will tell port * if agl-devel is on, then check journalctl * 05/12: no update * 05/26: Li did some more tests, issue/jira ? * Jan 24 14:49:47 raspberrypi3 afm-system-daemon[292]: ERROR: can't start system unit afm-appli-myapp--0.1--main@.service for uid 1001 [/w/workspace/release-jjb-halibut-snapshot/MACHINE/raspberrypi3/label/agl-test-slave/repoclone/output/tmp/work/armv7vet2hf-neon-vfpv4-agl-linux-gnueabi/af-main/halibut+gitAUTOINC+1289070816-r0/git/src/afm-urun.c:252] * 06/09: * no update * 06/23: * https://lava.automotivelinux.org/scheduler/job/10818#results_1029988 * Test wgts need inspection on issues: * FAIL: Failed to start application agl-service-geofence-test * ERROR: failed Resource temporarily unavailable * Update 05/26: WIP * Update 06/09: * https://jira.automotivelinux.org/browse/SPEC-3365 * No further updates * 06/23: * https://jira.automotivelinux.org/browse/SPEC-3365 * no updates * 07/07 * no update New: * Li: * test-wgt * https://jira.automotivelinux.org/browse/SPEC-2795 reopened -> devcall * https://jira.automotivelinux.org/browse/SPEC-3457 triage -> devcall * https://jira.automotivelinux.org/browse/SPEC-3466 triage -> devcall * working on missing tests and checking test coverage * Matthew and Brian * will try a pi4 configuration Next call: Tuesday July 21st - same timeslot. ==== Meeting 2020/06/23 ==== === Attendees === Jan-Simon, Scott, Brian, Matthew, Li ===Topics=== * infra: please try responsiveness of * https://confluence.automotivelinux.org/display/MAIN/AGL+Long+Term+Support * vs. * https://test-confluence.automotivelinux.org/display/TEST/AGL+Long+Term+Support * Fujitsu test report update (Fujitsu) * 06/09: * https://lists.automotivelinux.org/g/agl-dev-community/message/8413 * Pi4 no audio ... let's check driver module/kernel config and wireplumber-config device name string. * 06/23 * Pi4 no audio ... let's check driver module/kernel config and wireplumber-config device name * we do build CONFIG_SND_BCM2835=m * there is no wireplumber-config for the pi4, so the default might end-up only on hdmi (best guess) * https://git.automotivelinux.org/AGL/meta-agl-devel/tree/meta-pipewire/recipes-multimedia/wireplumber/wireplumber-board-config-agl?h=master * Fuego / Lava support * Update 12/10: * Reworking https://git.lavasoftware.org/lava/lava/merge_requests/330 * waiting for upstream review * Update 05/26 * ported and staged for 2020.05 * Update 06/09: * Corentin will ping Remi * Update 06/23 * No update from Corentin * Corentin: * Update 06/09: * Past week * kernelci-dev updated, new metadata in the jobs required * updated releng-scripts to python3 , JS needs to check CI builders before we merge * This week * blocked on python3 conversion * verification of kernelci- update with new metadata * Update 06/23 * Past week * transitioned releng to py3, builders were updated, but jinja2 version issue * https://gerrit.automotivelinux.org/gerrit/c/ci-management/+/24810 * This week * should have all deployed by EOW * Application tests , build multiple wgts * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * Update 03/03: * postponed after test wgt * 04/28: * Scott will post his first round of aglwgt/autobuild fixes after next is merged. * Plan is to review in gerrit this week * 05/12: * Scott has his patches for review in gerrit * https://gerrit.automotivelinux.org/gerrit/q/topic:%2522smurray/aglwgt%2522 * First patchset out of a series. * 05/26: * Next set of patches in preparation * 06/23: * aglwgt bbclass rewrite * waiting for marius branch to land first * AGL license scanning (SPDX/fossology) * 05/12: * License clearing done. * Next step? * 05/26 * Next ? * extend clearing work ? * more boards * or more packages * discussion about completing one target. Li will discuss further. * fossology has a 'reuse' decider that can built on existing decisions from previous runs. * 06/09: * 9.0.2 being uploaded to the fossology server * no analysis run, yet. still making sure all artifacts are fully uploaded. * found issue Dockerfile should install unzip (so we we don't use the php built-in), verifying that. Will have Ryan update our container. * 06/23: * unzip issue still a WIP item as: * during 'ununpack', we cannot abort the agent. But also cannot reupload the same file * the agents can get stuck -> unstable operation * SW360: * Will send Dockerfile for sw360 container * Update 03/03: * eta in 2 weeks * Update 03/31: * tests of sw360 in progress, working on 2 issues atm * Update 04/13: * sw360 Dockerfile and user guide:{{:eg-ciat:sw360.7z|here}}. * Todo Jan-Simon: setup sw360 with ryan * Setup server on same aws host but increase cpu/mem * Go from there. * Update 04/28: * Afraid we cannot use the Dockerfile as-is in AGL infra * Items to work on: * a) no proxy settings, no proxy user, no proxy pass * TLDR: if we deploy this in AWS, there is no Proxy at all. * b) Dockerfile uses downloaded Oracle jdk: can't we use openjdk installed in the distro installed by apt-get? * c) all data/databases need to be on persistent shared volume(s) * d) all setup needs to be automatic (in Dockerfile) * no manual setup steps - only docker build & docker run * e) ? would it make sense to split using docker-compose ? * Who can work on this ? @Kusakabe-san * Update 05/12: * work started on Dockerfile * Update 05/26: * a), b), c) completed * d) wip * let's exchange email * Update 06/09: * https://wiki.automotivelinux.org/_media/eg-ciat/sw360.zip new file * AI JS: Check with Ryan how to deploy * Update 06/23: * Still working through it with * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click * Update 06/09 * based on KernelCI instanced update * Li: how to test binding manually (post/token) * afaik afm-util show id will tell port * if agl-devel is on, then check journalctl * 05/12: no update * 05/26: Li did some more tests, issue/jira ? * Jan 24 14:49:47 raspberrypi3 afm-system-daemon[292]: ERROR: can't start system unit afm-appli-myapp--0.1--main@.service for uid 1001 [/w/workspace/release-jjb-halibut-snapshot/MACHINE/raspberrypi3/label/agl-test-slave/repoclone/output/tmp/work/armv7vet2hf-neon-vfpv4-agl-linux-gnueabi/af-main/halibut+gitAUTOINC+1289070816-r0/git/src/afm-urun.c:252] * 06/09: * no update * 06/23: * https://lava.automotivelinux.org/scheduler/job/10818#results_1029988 * Test wgts need inspection on issues: * FAIL: Failed to start application agl-service-geofence-test * ERROR: failed Resource temporarily unavailable * Update 05/26: WIP * Update 06/09: * https://jira.automotivelinux.org/browse/SPEC-3365 * No further updates * 06/23: * https://jira.automotivelinux.org/browse/SPEC-3365 * no updates New: * Li: starting point for (ui) app: * qmake example: https://git.automotivelinux.org/apps/dashboard/tree/ * cmake example: https://git.automotivelinux.org/apps/mixer/tree/ * working on review of LICENSE files in app repos * bbe: source meta-agl/scripts/aglsetup.sh -m bbe -b build-bb3 agl-demo agl-devel * bitbake agl-image-minimal * bitbake agl-demo-platform # afaik works with 720p Next call: Tuesday July 7th - same timeslot. ==== Meeting 2020/06/09 ==== === Attendees === Jan-Simon, Li, Scott, Corentin, Takada-san ===Topics=== * infra: please try responsiveness of * https://confluence.automotivelinux.org/display/MAIN/AGL+Long+Term+Support * vs. * https://test-confluence.automotivelinux.org/display/TEST/AGL+Long+Term+Support * Fujitsu test report update (Fujitsu) * 06/09: * https://lists.automotivelinux.org/g/agl-dev-community/message/8413 * Pi4 no audio ... let's check driver module/kernel config and wireplumber-config device name string. * Fuego / Lava support * Update 12/10: * Reworking https://git.lavasoftware.org/lava/lava/merge_requests/330 * waiting for upstream review * Update 05/12: * Test in progress * Test ported to 2020.04 * Update 05/26 * ported and staged for 2020.05 * Update 06/09: * Corentin will ping Remi * Corentin: * Update 05/26 * Past week * Pi4 no change, u-boot master did not help * testing the patch and upstreaming * This week * email reports through kernelci.automotivelinux.org - prepare uprev * Update 06/09: * Past week * kernelci-dev updated, new metadata in the jobs required * updated releng-scripts to python3 , JS needs to check CI builders before we merge * This week * blocked on python3 conversion * verification of kernelci- update with new metadata * Application tests , build multiple wgts * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * Update 03/03: * postponed after test wgt * 04/28: * Scott will post his first round of aglwgt/autobuild fixes after next is merged. * Plan is to review in gerrit this week * 05/12: * Scott has his patches for review in gerrit * https://gerrit.automotivelinux.org/gerrit/q/topic:%2522smurray/aglwgt%2522 * First patchset out of a series. * 05/26: * Next set of patches in preparation * AGL license scanning (SPDX/fossology) * 05/12: * License clearing done. * Next step? * 05/26 * Next ? * extend clearing work ? * more boards * or more packages * discussion about completing one target. Li will discuss further. * fossology has a 'reuse' decider that can built on existing decisions from previous runs. * 06/09: * 9.0.2 being uploaded to the fossology server * no analysis run, yet. still making sure all artifacts are fully uploaded. * found issue Dockerfile should install unzip (so we we don't use the php built-in), verifying that. Will have Ryan update our container. * SW360: * Will send Dockerfile for sw360 container * Update 03/03: * eta in 2 weeks * Update 03/31: * tests of sw360 in progress, working on 2 issues atm * Update 04/13: * sw360 Dockerfile and user guide:{{:eg-ciat:sw360.7z|here}}. * Todo Jan-Simon: setup sw360 with ryan * Setup server on same aws host but increase cpu/mem * Go from there. * Update 04/28: * Afraid we cannot use the Dockerfile as-is in AGL infra * Items to work on: * a) no proxy settings, no proxy user, no proxy pass * TLDR: if we deploy this in AWS, there is no Proxy at all. * b) Dockerfile uses downloaded Oracle jdk: can't we use openjdk installed in the distro installed by apt-get? * c) all data/databases need to be on persistent shared volume(s) * d) all setup needs to be automatic (in Dockerfile) * no manual setup steps - only docker build & docker run * e) ? would it make sense to split using docker-compose ? * Who can work on this ? @Kusakabe-san * Update 05/12: * work started on Dockerfile * Update 05/26: * a), b), c) completed * d) wip * let's exchange email * Update 06/09: * https://wiki.automotivelinux.org/_media/eg-ciat/sw360.zip new file * AI JS: Check with Ryan how to deploy * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click * Update 05/12 * WIP * Update 05/26 * no update * Update 06/09 * based on KernelCI instanced update * Li: how to test binding manually (post/token) * afaik afm-util show id will tell port * if agl-devel is on, then check journalctl * 05/12: no update * 05/26: Li did some more tests, issue/jira ? * Jan 24 14:49:47 raspberrypi3 afm-system-daemon[292]: ERROR: can't start system unit afm-appli-myapp--0.1--main@.service for uid 1001 [/w/workspace/release-jjb-halibut-snapshot/MACHINE/raspberrypi3/label/agl-test-slave/repoclone/output/tmp/work/armv7vet2hf-neon-vfpv4-agl-linux-gnueabi/af-main/halibut+gitAUTOINC+1289070816-r0/git/src/afm-urun.c:252] * 06/09: * no update * Test wgts need inspection on issues: * FAIL: Failed to start application agl-service-geofence-test * ERROR: failed Resource temporarily unavailable * Update 05/26: WIP * Update 06/09: * https://jira.automotivelinux.org/browse/SPEC-3365 * No further updates New: * Next call: Tuesday June 23th - same timeslot. ------------ ==== Meeting 2020/05/26 ==== === Attendees === Jan-Simon, Walt, Corentin, Li, Takada-san ===Topics=== * infra: please send me the output of: (jira/gerrit/lava) * mtr -4 -rnw -c 30 gerrit.automotivelinux.org * mtr -4 -rnw -c 30 ec2-54-209-7-50.compute-1.amazonaws.com * Fujitsu test report update (Fujitsu) * 05/26: * No update * Next release is 9.0.2 eta 2020-06-05 * Fuego / Lava support * Update 12/10: * Reworking https://git.lavasoftware.org/lava/lava/merge_requests/330 * waiting for upstream review * Update 05/12: * Test in progress * Test ported to 2020.04 * Update 05/26 * ported and staged for 2020.05 * Corentin: * Update 05/12 * Past week: * Deployed lava-docker 2020.02 and fixed issues * pi4 has still issues with network/tftp timeouts * This week: * Fix h3 - currently h3 has issues with switch deployed to fix pi4. Will revert to old switch. Let's file a jira for the pi4/switch issue. * Testing for https://git.lavasoftware.org/lava/lava/merge_requests/330 * ported and being tested * Update 05/26 * Past week * Pi4 no change, u-boot master did not help * testing the patch and upstreaming * This week * email reports through kernelci.automotivelinux.org - prepare uprev * Application tests , build multiple wgts * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * Update 03/03: * postponed after test wgt * 04/28: * Scott will post his first round of aglwgt/autobuild fixes after next is merged. * Plan is to review in gerrit this week * 05/12: * Scott has his patches for review in gerrit * https://gerrit.automotivelinux.org/gerrit/q/topic:%2522smurray/aglwgt%2522 * First patchset out of a series. * 05/26: * Next set of patches in preparation * AGL license scanning (SPDX/fossology) * 05/12: * License clearing done. * Next step? * 05/26 * Next ? * extend clearing work ? * more boards * or more packages * discussion about completing one target. Li will discuss further. * fossology has a 'reuse' decider that can built on existing decisions from previous runs. * SW360: * Will send Dockerfile for sw360 container * Update 03/03: * eta in 2 weeks * Update 03/31: * tests of sw360 in progress, working on 2 issues atm * Update 04/13: * sw360 Dockerfile and user guide:{{:eg-ciat:sw360.7z|here}}. * Todo Jan-Simon: setup sw360 with ryan * Setup server on same aws host but increase cpu/mem * Go from there. * Update 04/28: * Afraid we cannot use the Dockerfile as-is in AGL infra * Items to work on: * a) no proxy settings, no proxy user, no proxy pass * TLDR: if we deploy this in AWS, there is no Proxy at all. * b) Dockerfile uses downloaded Oracle jdk: can't we use openjdk installed in the distro installed by apt-get? * c) all data/databases need to be on persistent shared volume(s) * d) all setup needs to be automatic (in Dockerfile) * no manual setup steps - only docker build & docker run * e) ? would it make sense to split using docker-compose ? * Who can work on this ? @Kusakabe-san * Update 05/12: * work started on Dockerfile * Update 05/26: * a), b), c) completed * d) wip * let's exchange email * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click * Update 05/12 * WIP * Update 05/26 * no update * Li: how to test binding manually (post/token) * afaik afm-util show id will tell port * if agl-devel is on, then check journalctl * 05/12: no update * 05/26: Li did some more tests, issue/jira ? * Jan 24 14:49:47 raspberrypi3 afm-system-daemon[292]: ERROR: can't start system unit afm-appli-myapp--0.1--main@.service for uid 1001 [/w/workspace/release-jjb-halibut-snapshot/MACHINE/raspberrypi3/label/agl-test-slave/repoclone/output/tmp/work/armv7vet2hf-neon-vfpv4-agl-linux-gnueabi/af-main/halibut+gitAUTOINC+1289070816-r0/git/src/afm-urun.c:252] * Test wgts need inspection on issues: * FAIL: Failed to start application agl-service-geofence-test * ERROR: failed Resource temporarily unavailable * Update 05/26: WIP New: * Next call: Tuesday June 9th - same timeslot. -------- ==== Meeting 2020/05/12 ==== === Attendees === Jan-Simon, Walt, Li, Corentin, Farshid, ===Topics=== * infra: please send me the output of: * mtr -4 -rw -c 30 gerrit.automotivelinux.org * mtr -4 -rw -c 30 ec2-54-209-7-50.compute-1.amazonaws.com * Fujitsu test report update (Fujitsu) * 04/28: * Not this week * 05/12: * Not this week * Fuego / Lava support * Update 12/10: * Reworking https://git.lavasoftware.org/lava/lava/merge_requests/330 * waiting for upstream review * Update 03/31: * Corentin to push for a review upstream * Update 04/28: * Upstream requested testing (-> Corentin) * Update 05/12: * Test in progress * Test ported to 2020.04 * Corentin: * Update 04/28 * Past week: * Fixed Pi issues (basically replacing the network switch -> likely autonegotiation issue) * This week: * Finish lava-docker * Update 05/12 * Past week: * Deployed lava-docker 2020.02 and fixed issues * pi4 has still issues with network/tftp timeouts * This week: * Fix h3 - currently h3 has issues with switch deployed to fix pi4. Will revert to old switch. Let's file a jira for the pi4/switch issue. * Testing for https://git.lavasoftware.org/lava/lava/merge_requests/330 * ported and being tested * Application tests , build multiple wgts * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * 03/31: * autobuild issue: https://jira.automotivelinux.org/browse/SPEC-3300 * 04/14: * Scott is working on a fix of the autobuild issue. * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * Update 03/03: * postponed after test wgt * 04/28: * Scott will post his first round of aglwgt/autobuild fixes after next is merged. * Plan is to review in gerrit this week * 05/12: * Scott has his patches for review in gerrit * https://gerrit.automotivelinux.org/gerrit/q/topic:%2522smurray/aglwgt%2522 * First patchset out of a series. * AGL license scanning (SPDX/fossology) * Update 04/28: * Li will review these packages in Fossology: * af-binder * af-binder-devtools-native * afb-test * af-main * af-platform-setup * agl-audio-plugin * agl-compositor * agl-demo-platform.bb * agl-desktop-config * agl-login-manager * agl-service-audiomixer * agl-service-bluetooth * agl-service-bluetooth-map * agl-service-bluetooth-pbap * agl-service-can-high-level-viwi * agl-service-can-low-level * agl-service-data-persistence * agl-service-geoclue * agl-service-geofence * agl-service-gps * agl-service-homescreen * agl-service-hvac * agl-service-hvac-conf * agl-service-identity-agent * agl-service-iiodevices * agl-service-mediaplayer * agl-service-mediascanner * agl-service-navigation * agl-service-network * agl-service-nfc * agl-service-platform-info * agl-service-radio * agl-service-signal-composer * agl-service-steering-wheel * agl-service-taskmanager * agl-service-telephony * agl-service-unicens * agl-service-unicens-controller * agl-service-voice-high * agl-service-voice-high-capabilities * agl-service-weather * agl-service-windowmanager * agl-users * alexa-viewer * alexa-voiceagent-config * alexa-voiceagent-service * alexa-voice-high-config * als-meter-demo * audit * bluez-alsa * btwilink-disable-conf * cmake-apps-module * connman * connman-ncurses * controls * cynagoauth * cynagora * cynagora-cynara-compat * dashboard * default-voice-high-config * demo-i2c-udev-conf * dev-mapping * dev-mapping-demo * flite * flite-voicedata * high-level-viwi-service.bb * hmi-debug * homescreen * html5-dashboard * html5-homescreen * html5-hvac * html5-launcher * html5-mediaplayer * html5-mixer * html5-settings * hts-engine * hvac * initramfs-netboot * inputeventmanager * keyutils * launcher * libafb-helpers * libafb-helpers-qt * libappcontroller * libhomescreen * libmicrohttpd * libmp4v2 * libnfc * libqtappfw * libwindowmanager * libzip * lightmediascanner * lin-config * low-can-demo * low-level-can-generator * mediaplayer * messaging * mixer * most.bb * nativesdk-af-binder-devtools * nativesdk-af-main * navigation-demo * neardal * nghttp2 * noto-emoji * nss-localuser * ondemandnavi * ondemandnavi-config * onscreenapp * openjtalk * openjtalk-voicedata * osmium * osrm-backend * phone * pipewire * pipewire-conf-agl * poiapp * poiapp-api-key * protozero * qlibhomescreen * qlibwindowmanager * qml-execscript-plugin * qtaglextras * qtcompositor-conf * qt-qrcode * qtquickcontrols2-agl * qtquickcontrols2-agl-style * radio * rtl-sdr * run-agl-postinsts * runxdg * security-manager * settings * settings-log-utils * sllin * smacknet * smack-system-setup * systemd-agl-sync * taskmanager * tbtnavi * unicens-config * vboxguestdrivers * vod-server * waltham * waltham-transmitter * wam * wam-tinyproxy * wayland-ivi-extension * weston-ready * windowmanager * wireplumber * wireplumber-board-config-agl * 05/12: * License clearing done. * Next step? * SW360: * Will send Dockerfile for sw360 container * Update 03/03: * eta in 2 weeks * Update 03/31: * tests of sw360 in progress, working on 2 issues atm * Update 04/13: * sw360 Dockerfile and user guide:{{:eg-ciat:sw360.7z|here}}. * Todo Jan-Simon: setup sw360 with ryan * Setup server on same aws host but increase cpu/mem * Go from there. * Update 04/28: * Afraid we cannot use the Dockerfile as-is in AGL infra * Items to work on: * a) no proxy settings, no proxy user, no proxy pass * TLDR: if we deploy this in AWS, there is no Proxy at all. * b) Dockerfile uses downloaded Oracle jdk: can't we use openjdk installed in the distro installed by apt-get? * c) all data/databases need to be on persistent shared volume(s) * d) all setup needs to be automatic (in Dockerfile) * no manual setup steps - only docker build & docker run * e) ? would it make sense to split using docker-compose ? * Who can work on this ? @Kusakabe-san * Update 05/12: * work started on Dockerfile * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click * Update 05/12 * WIP * Li: how to test binding manually (post/token) * afaik afm-util show id will tell port * if agl-devel is on, then check journalctl * 03/31: no update * 04/14: no update * 04/28: no update * 05/12: no update New: * Test wgts need inspection on issues: * FAIL: Failed to start application agl-service-geofence-test * ERROR: failed Resource temporarily unavailable *Li: during license clearing: will review packages with non-apacheV2 in AGL. Next call: Tuesday May 26th - same timeslot. ==== Meeting 2020/04/28 ==== === Attendees === Jan-Simon, Walt, Corentin, Li, Parth, Takada-san ===Topics=== * Fujitsu test report update (Fujitsu) * 04/28: * Not this week * Fuego / Lava support * Update 12/10: * Reworking https://git.lavasoftware.org/lava/lava/merge_requests/330 * waiting for upstream review * Update 03/31: * Corentin to push for a review upstream * Update 04/28: * Upstream requested testing (-> Corentin) * Corentin: * Update 04/14 * Past week: * lava uprev * pi4 passes health check: https://lava.automotivelinux.org/scheduler/job/9999 * This week: * schedule lava uprev w/ Ryan * Update 04/28 * Past week: * Fixed Pi issues (basically replacing the network switch -> likely autonegotiation issue) * This week: * Finish lava-docker * Application tests , build multiple wgts * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * 03/31: * autobuild issue: https://jira.automotivelinux.org/browse/SPEC-3300 * 04/14: * Scott is working on a fix of the autobuild issue. * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * Update 03/03: * postponed after test wgt * 04/28: * Scott will post his first round of aglwgt/autobuild fixes after next is merged. * Plan is to review in gerrit this week * AGL license scanning (SPDX/fossology) * Update 04/14: * 9.0.1 uploaded, ready for review - Li will review 9.0.1 * Update 04/28: * Li will review these packages in Fossology: * af-binder * af-binder-devtools-native * afb-test * af-main * af-platform-setup * agl-audio-plugin * agl-compositor * agl-demo-platform.bb * agl-desktop-config * agl-login-manager * agl-service-audiomixer * agl-service-bluetooth * agl-service-bluetooth-map * agl-service-bluetooth-pbap * agl-service-can-high-level-viwi * agl-service-can-low-level * agl-service-data-persistence * agl-service-geoclue * agl-service-geofence * agl-service-gps * agl-service-homescreen * agl-service-hvac * agl-service-hvac-conf * agl-service-identity-agent * agl-service-iiodevices * agl-service-mediaplayer * agl-service-mediascanner * agl-service-navigation * agl-service-network * agl-service-nfc * agl-service-platform-info * agl-service-radio * agl-service-signal-composer * agl-service-steering-wheel * agl-service-taskmanager * agl-service-telephony * agl-service-unicens * agl-service-unicens-controller * agl-service-voice-high * agl-service-voice-high-capabilities * agl-service-weather * agl-service-windowmanager * agl-users * alexa-viewer * alexa-voiceagent-config * alexa-voiceagent-service * alexa-voice-high-config * als-meter-demo * audit * bluez-alsa * btwilink-disable-conf * cmake-apps-module * connman * connman-ncurses * controls * cynagoauth * cynagora * cynagora-cynara-compat * dashboard * default-voice-high-config * demo-i2c-udev-conf * dev-mapping * dev-mapping-demo * flite * flite-voicedata * high-level-viwi-service.bb * hmi-debug * homescreen * html5-dashboard * html5-homescreen * html5-hvac * html5-launcher * html5-mediaplayer * html5-mixer * html5-settings * hts-engine * hvac * initramfs-netboot * inputeventmanager * keyutils * launcher * libafb-helpers * libafb-helpers-qt * libappcontroller * libhomescreen * libmicrohttpd * libmp4v2 * libnfc * libqtappfw * libwindowmanager * libzip * lightmediascanner * lin-config * low-can-demo * low-level-can-generator * mediaplayer * messaging * mixer * most.bb * nativesdk-af-binder-devtools * nativesdk-af-main * navigation-demo * neardal * nghttp2 * noto-emoji * nss-localuser * ondemandnavi * ondemandnavi-config * onscreenapp * openjtalk * openjtalk-voicedata * osmium * osrm-backend * phone * pipewire * pipewire-conf-agl * poiapp * poiapp-api-key * protozero * qlibhomescreen * qlibwindowmanager * qml-execscript-plugin * qtaglextras * qtcompositor-conf * qt-qrcode * qtquickcontrols2-agl * qtquickcontrols2-agl-style * radio * rtl-sdr * run-agl-postinsts * runxdg * security-manager * settings * settings-log-utils * sllin * smacknet * smack-system-setup * systemd-agl-sync * taskmanager * tbtnavi * unicens-config * vboxguestdrivers * vod-server * waltham * waltham-transmitter * wam * wam-tinyproxy * wayland-ivi-extension * weston-ready * windowmanager * wireplumber * wireplumber-board-config-agl * SW360: * Will send dockerfile for sw360 container * Update 03/03: * eta in 2 weeks * Update 03/31: * tests of sw360 in progress, working on 2 issues atm * Update 04/13: * sw360 dockerfile and user guide:{{:eg-ciat:sw360.7z|here}}. * Todo Jan-Simon: setup sw360 with ryan * Setup server on same aws host but increase cpu/mem * Go from there. * Update 04/28: * Afraid we cannot use the Dockerfile as-is in AGL infra * Items to work on: * a) no proxy settings, no proxy user, no proxy pass * TLDR: if we deploy this in AWS, there is no Proxy at all. * b) Dockerfile uses downloaded Oracle jdk: can't we use openjdk installed in the distro installed by apt-get? * c) all data/databases need to be on persistent shared volume(s) * d) all setup needs to be automatic (in Dockerfile) * no manual setup steps - only docker build & docker run * e) ? would it make sense to split using docker-compose ? * Who can work on this ? @Kusakabe-san * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click * Update 04/14 * No update * Update 04/28 * No update * Li: how to test binding manually (post/token) * afaik afm-util show id will tell port * if agl-devel is on, then check journalctl * 03/31: no update * 04/14: no update * 04/28: no update * Li: * Fossology: reported need to restart container. we did not have to restart fossology.automotivelinux.org, yet. * 04/28: on fossology.automotivelinux.org scheduler was stuck on one job New: * Li: * pi4 image - pi4 image * $ source meta-agl/scripts/aglsetup.sh -m raspberrypi4 agl-demo agl-netboot agl-appfw-smack * $ source meta-agl/scripts/aglsetup.sh -m raspberrypi4 agl-demo >agl-devel< * Next call: Tuesday May 12th - same timeslot. ==== Meeting 2020/04/14 ==== === Attendees === Jan-Simon, Walt, Kusakabe-san, Li Xiaoming, Kurokawa-san ===Topics=== * Fujitsu test report update (Fujitsu) * 04/14: * 9.0.1: still todo * Fuego / Lava support * Update 12/10: * Reworking https://git.lavasoftware.org/lava/lava/merge_requests/330 * waiting for upstream review * Update 03/31: * Corentin to push for a review upstream * Corentin: * Update 03/31 * Past week: * Testing new lava version in baylibre lab * allow multiple artifacts to be uploaded (combined qemu+intel case) * This week: * prepare lava uprev * pi4 in corelab * Update 04/14 * Past week: * lava uprev * pi4 passes health check: https://lava.automotivelinux.org/scheduler/job/9999 * This week: * schedule lava uprev w/ Ryan * Application tests , build multiple wgts * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * 03/31: * autobuild issue: https://jira.automotivelinux.org/browse/SPEC-3300 * 04/14: * Scott is working on a fix of the autobuild issue. * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * Update 03/03: * postponed after test wgt * * AGL license scanning (SPDX/fossology) * Update 31/03: * Jan-Simon: rewrote the fossology-rest class into a phased version to * allow upload / analysis / spdx report to be separate steps * this works with online servers in remote datacenters with limited CPU * Update 04/14: * 9.0.1 uploaded, ready for review - Li will review 9.0.1 * Kusakabe-san: * Will send dockerfile for sw360 container * Update 03/03: * eta in 2 weeks * Update 03/31: * tests of sw360 in progress, working on 2 issues atm * Update 04/13: * sw360 dockerfile and user guide:{{:eg-ciat:sw360.7z|here}}. * Todo Jan-Simon: setup sw360 with ryan * Setup server on same aws host but increase cpu/mem * Go from there. * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click * Update 03/31: * No update * Update 04/14 * No update * CI Setup document needed - https://jira.automotivelinux.org/browse/SPEC-2980 * Update 12/10: * Files collected. Markdown changes pending. * Update 03/03: * Tbd (AMM cancelled) * Update 04/14: * preparing a presentation * Li: how to test binding manually (post/token) * afaik afm-util show id will tell port * if agl-devel is on, then check journalctl * 03/31: no update * 04/14: no update New: * Li: * Fossology: reported need to restart container. we did not have to restart fossology.automotivelinux.org, yet. * Next call: Tuesday Apr 28th - same timeslot. -------- ==== Meeting 2020/03/31 ==== === Attendees === Jan-Simon, Li Xiaoming, Walt, Parth, Kusakabe-san, Corentin ===Topics=== * Fujitsu test report update (Fujitsu) * 03/30: 8.0.6 test report * https://lists.automotivelinux.org/g/agl-dev-community/message/8185 * https://jira.automotivelinux.org/browse/SPEC-2551 and https://jira.automotivelinux.org/browse/SPEC-2620 reopen if it applies to master * Fuego / Lava support * Update 12/10: * Reworking https://git.lavasoftware.org/lava/lava/merge_requests/330 * waiting for upstream review * Update 03/03: * Who can drive this upstream? Corentin ? * Update 03/31: * Corentin to push for a review upstream * Corentin: * Update 03/03: * Past week: * Pi4 is now working https://lava.automotivelinux.org/scheduler/job/10209 * This week: * Deal with enabling in CI * Also fix board boot issues after zeus uprev * https://jira.automotivelinux.org/browse/SPEC-3221 * Update 03/31 * Past week: * Testing new lava version in baylibre lab * allow multiple artifacts to be uploaded (combined qemu+intel case) * This week: * prepare lava uprev * pi4 in corelab * Application tests , build multiple wgts * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * 03/31: * autobuild issue: https://jira.automotivelinux.org/browse/SPEC-3300 * AGL license scanning (SPDX/fossology) * Update 02/18: * Debugging python error * Uploading a build of icefish 9.0.0 to fossology (qemux86-64) * AI: create user for Li * AI: check if m3/h3-nogfx can be uploaded * Update 03/03: * Still stuck on python error see: * https://jira.automotivelinux.org/browse/SPEC-2960 * Update 31/03: * Jan-Simon: rewrote the fossology-rest class into a phased version to * allow upload / analysis / spdx report to be separate steps * this works with online servers in remote datacenters with limited CPU * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * Update 03/03: * postponed after test wgt * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click * Update 12/10: * Please comment on https://jira.automotivelinux.org/browse/SPEC-3020 * Update 01/21: * New fields being added for the links back (https://gerrit.automotivelinux.org/gerrit/#/c/AGL/releng-scripts/+/23580) * Update 03/03: * No update * Update 03/31: * No update * CI Setup document needed - https://jira.automotivelinux.org/browse/SPEC-2980 * Update 12/10: * Files collected. Markdown changes pending. * Update 03/03: * Tbd (AMM cancelled) * Update 03/31: * tbd * Li: how to test binding manually (post/token) * afaik afm-util show id will tell port * if agl-devel is on, then check journalctl * 03/31: no update * Kusakabe-san: * Will send dockerfile for sw360 container * Update 03/03: * eta in 2 weeks * Update 03/31: * tests of sw360 in progress, working on 2 issues atm * Update 04/13: * sw360 dockerfile and user guide:{{:eg-ciat:sw360.7z|here}}. Next call: Tuesday Apr 14th - same timeslot. -------- ==== Meeting 2020/03/03 ==== === Attendees === Jan-Simon, Walt, Michail Zaytsev (Mera), Andrey Shamanin (Mera), Leonid Lazarev (Mera), Marat (Mera), Li, Ronan, Frederic, Corentin, Takada-san, ===Topics=== * Fujitsu test report update (Fujitsu) * 02/18: * 9.0.0 report sent: https://lists.automotivelinux.org/g/agl-dev-community/message/8120 * https://jira.automotivelinux.org/browse/SPEC-2305 * https://jira.automotivelinux.org/browse/SPEC-2811 * https://jira.automotivelinux.org/browse/SPEC-2666 -> Known issue * Fuego / Lava support * Update 12/10: * Reworking https://git.lavasoftware.org/lava/lava/merge_requests/330 * waiting for upstream review * Update 02/18: * No update from upstream * Update 03/03: * Who can drive this upstream? Corentin ? * Corentin: * Update 02/18: * Corentin nbd/netbooted the rpi4 via using the CONFIG_ARCH_FIXUP_FDT_MEMORY of uboot. * Good news, we do not need bootefi, so the boot process is classic. * Board to be added to lab asap * Update 03/03: * Past week: * Pi4 is now working https://lava.automotivelinux.org/scheduler/job/10209 * This week: * Deal with enabling in CI * Also fix board boot issues after zeus uprev * https://jira.automotivelinux.org/browse/SPEC-3221 * Application tests , build multiple wgts * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * Update 11/26: * Fred: working on cmake module update, mods in sandbox (https://git.automotivelinux.org/src/cmake-apps-module/commit/?h=sandbox/DDTLK/autobuild&id=0c42794f11bc2d6a0d2d14b3a57f5222a3418b6a) * Move to master and test on app builds * Update 02/18: * New patches in gerrit topic:SPEC-2049 * https://gerrit.automotivelinux.org/gerrit/#/q/topic:SPEC-2049 * Update 03/03: * Need to deinig submodule and rebase https://gerrit.automotivelinux.org/gerrit/#/c/apps/agl-service-can-high-level-viwi/+/24015/ * AGL license scanning (SPDX/fossology) * Update 02/18: * Debugging python error * Uploading a build of icefish 9.0.0 to fossology (qemux86-64) * AI: create user for Li * AI: check if m3/h3-nogfx can be uploaded * Update 03/03: * Still stuck on python error see: * https://jira.automotivelinux.org/browse/SPEC-2960 * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * Update 03/03: * postponed after test wgt * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click * Update 12/10: * Please comment on https://jira.automotivelinux.org/browse/SPEC-3020 * Update 01/21: * New fields being added for the links back (https://gerrit.automotivelinux.org/gerrit/#/c/AGL/releng-scripts/+/23580) * Update 03/03: * No update * CI Setup document needed - https://jira.automotivelinux.org/browse/SPEC-2980 * Update 12/10: * Files collected. Markdown changes pending. * Update 03/03: * Tbd (AMM cancelled) * Li: how to test binding manually (post/token) * afaik afm-util show id will tell port * if agl-devel is on, then check journalctl * Update 03/03: * do we have enough documentation? * from irc * ''11:14 jobol: Hello, I have a question about the appfw token. In a binding, the server claims a token. If the client do not provide the same token, the connection can not be established. Right ? 11:16 https://docs.automotivelinux.org/docs/en/master/apis_services/reference/af-binder/afb-binding-writing.html , example tuto-1.c 11:17 LiXiaoming, the connection can be established but it lakes something 11:17 This is true for HTTP requests 11:19 When switching to websockets, the token is required or not depending on the version 11:19 management of tokens changed from HALIBUT to ICEFISH 11:20 Now I am testing on native Ubuntu 18.04. Both http and websocket. I found token is meaningless here. 11:21 On latest versions yes 11:21 But it depends if you have the cynagora backend 11:22 yeah, I skip the cynagora install. 11:22 Normally on AGL the binder contacts the permission database to validate the token 11:23 Cynagora can be installed on UBUNTU 11:23 SPEC-3212. Because I met issue when install Cynagora 11:23 I know you had issue in installing cynara but that is not the same 11:24 cynara isn't cynagora 11:24 Are you compiling the binder? 11:24 on ubuntu? 11:24 '=( . Ubuntu 18.0.4 11:24 yes . I am compiling and install binder on Ubuntu 11:25 if cynagora (https://git.automotivelinux.org/src/cynagora/) doesn't compile on ubuntu, let me know. 11:26 But for samsung's cynara i'm not going to help 11:26 cynagora provide a compatibility library for cynara 11:27 I'm working on a binder 100% cynagora. 11:28 I can push a pre-release branch if you want to check it and help to make it work 11:29 just compile and install cynagora is enough ? Do I any need other steps to set cynagora as backend ？ 11:30 this is enough to active permission check. Then when active, the service should run (cynagorad) 11:31 on my computer I'm using the prefix ~/.local for my installations but the default is to use the prefix /usr/local 11:38 Good to know that. I will try it with cynagora's latest branch again. Thank you.'' * Kusakabe-san: * Will send dockerfile for sw360 container * Update 03/03: * eta in 2 weeks New: * Next call: Tuesday Mar 31st - same timeslot. -------- ==== Meeting 2020/02/18 ==== === Attendees === Jan-Simon, Ronan, Li, Frederic, Kusakabe-san ===Topics=== * Fujitsu test report update (Fujitsu) * 02/18: * 8.99.5 to be sent. * Fuego / Lava support * Update 12/10: * Reworking https://git.lavasoftware.org/lava/lava/merge_requests/330 * waiting for upstream review * Update 02/18: * No update from upstream * Corentin: * Update 02/06: * Pi4 does not boot with booti (64bit) on u-boot with initrd. Alternative boot with efi does not take initrd. * Update 02/18: * Corentin nbd/netbooted the rpi4 via using the CONFIG_ARCH_FIXUP_FDT_MEMORY of uboot. * Good news, we do not need bootefi, so the boot process is classic. * Board to be added to lab asap * Application tests , build multiple wgts * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * Update 11/26: * Fred: working on cmake module update, mods in sandbox (https://git.automotivelinux.org/src/cmake-apps-module/commit/?h=sandbox/DDTLK/autobuild&id=0c42794f11bc2d6a0d2d14b3a57f5222a3418b6a) * Move to master and test on app builds * Update 01/21: * native build works, issues on the yocto build * reconfiguration issues * proposal use separate build folders for each variant (normal, debug, test, coverage) * Update 02/06: * No update yet on https://jira.automotivelinux.org/browse/SPEC-2049 and application repo changes pending. * Update 02/18: * New patches in gerrit topic:SPEC-2049 * https://gerrit.automotivelinux.org/gerrit/#/q/topic:SPEC-2049 * AGL license scanning (SPDX/fossology) * Update 02/06: * Updated meta layer in gerrit * Server ready https://fossology.automotivelinux.org/repo/ * Icefish to be build with upload to fossology (just qemux86-64) * Update 02/18: * Debugging python error * Uploading a build of icefish 9.0.0 to fossology (qemux86-64) * AI: create user for Li * AI: check if m3/h3-nogfx can be uploaded * Li: wgt tests only work in guppy and not in halibut * https://jira.automotivelinux.org/browse/SPEC-2795 * Will do app tests on guppy until this is fixed * Update 11/14: * waiting for update from Fred on https://jira.automotivelinux.org/browse/SPEC-2795 * Update 11/26: * Fred will cherry-pick the needed changes to * Update 12/10: * No update, added Jose. * Update 01/21: * SPEC-2795 - Stephane looking into it. * Update 02/06: * No update * Update 02/18: * https://jira.automotivelinux.org/browse/SPEC-2795 resolved * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * Update 01/21: * Update 02/06: * Update 02/18: * postponed after test wgt * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click * Update 12/10: * Please comment on https://jira.automotivelinux.org/browse/SPEC-3020 * Update 01/21: * New fields being added for the links back (https://gerrit.automotivelinux.org/gerrit/#/c/AGL/releng-scripts/+/23580) * Update 02/18: * No update * CI Setup document needed - https://jira.automotivelinux.org/browse/SPEC-2980 * Update 12/10: * Files collected. Markdown changes pending. * Update 02/18: * Presentation during AMM to be added to doc site. New: * Li: how to test binding manually (post/token) * afaik afm-util show id will tell port * if agl-devel is on, then check journalctl * Kusakabe-san: * Will send dockerfile for sw360 container Next call: Tuesday Mar 3rd - same timeslot. -------- ==== Meeting 2020/02/06 ==== === Attendees === Jan-Simon, Kusakabe-san ===Topics=== * Roadmap items and review for 2020 * Improved publishing of test results * Email Reports * improve format and feedback to gerrit * Additional tests * Applications * build and test different wgt types (normal, debug, test, coverage) * BSPs * KSelfTest * Fuego integration * Support to run tests as lava jobs (from within Fuego) * Images in build matrix * more variants ? qemu only ? * e.g. sota enabled * cluster profile/image * Fujitsu test report update (Fujitsu) * 8.0.5 * Update 02/06: Fujitsu-CN closed this week * Fuego / Lava support * Update 12/10: * Reworking https://git.lavasoftware.org/lava/lava/merge_requests/330 * waiting for upstream review * Update 01/21: * Update 02/06: * No update from upstream * Corentin: * Update 01/21 * Working on Pi4 and lab integration (network support in u-boot tested w/ upstream patch) * Fixing lab and adding lava-bot * Fixes for the vayu * Update 02/06: * Pi4 does not boot with booti (64bit) on u-boot with initrd. Alternative boot with efi does not take initrd. * Application tests , build multiple wgts * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * Update 11/26: * Fred: working on cmake module update, mods in sandbox (https://git.automotivelinux.org/src/cmake-apps-module/commit/?h=sandbox/DDTLK/autobuild&id=0c42794f11bc2d6a0d2d14b3a57f5222a3418b6a) * Move to master and test on app builds * Update 01/21: * native build works, issues on the yocto build * reconfiguration issues * proposal use separate build folders for each variant (normal, debug, test, coverage) * Update 02/06: * No update yet on https://jira.automotivelinux.org/browse/SPEC-2049 and application repo changes pending. * AGL license scanning (SPDX/fossology) * Update 10/15: * no update * http://95.216.96.43:8081/repo/?mod=browse is for qemux86-64 AGL HH v8.0.0 * Update 12/10: * SC approved server. * setting up fossology 3.6 (from docker) * Update 01/21: * IT brought up the instance in aws * Configuration start * https://github.com/dl9pf/meta-spdxscanner/commit/08a1e49cb97aae1033c6bd06ce29caf6c1df67dd * provides mechanism to do folder per release * Target: try with II rc5 * Update 02/06: * Updated meta layer in gerrit * Server ready https://fossology.automotivelinux.org/repo/ * Icefish to be build with upload to fossology (just qemux86-64) * Li: wgt tests only work in guppy and not in halibut * https://jira.automotivelinux.org/browse/SPEC-2795 * Will do app tests on guppy until this is fixed * Update 10/15: * will retest on master * Update 11/14: * waiting for update from Fred on https://jira.automotivelinux.org/browse/SPEC-2795 * Update 11/26: * Fred will cherry-pick the needed changes to * Update 12/10: * No update, added Jose. * Update 01/21: * SPEC-2795 - Stephane looking into it. * Update 02/06: * No update * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * Update 01/21: * Update 02/06: * postponed after test wgt * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click * Update 12/10: * Please comment on https://jira.automotivelinux.org/browse/SPEC-3020 * Update 01/21: * New fields being added for the links back (https://gerrit.automotivelinux.org/gerrit/#/c/AGL/releng-scripts/+/23580) * CI Setup document needed - https://jira.automotivelinux.org/browse/SPEC-2980 * Update 12/10: * Files collected. Markdown changes pending. * Update 01/21: * Update 02/06: * No update New: * Next call: Tuesday Feb 18th - same timeslot. -------- ==== Meeting 2019/12/10 ==== === Attendees === Jan-Simon, Fred, Li, Takada-san, Stephane, Walt ===Topics=== * Fujitsu test report update (Fujitsu) * Update 11/26: * Bugs filed after testing 8.99.1 / RC1 * https://jira.automotivelinux.org/browse/SPEC-2986 * https://jira.automotivelinux.org/browse/SPEC-2987 * https://jira.automotivelinux.org/browse/SPEC-2712 (reopen) * https://jira.automotivelinux.org/browse/SPEC-2630 (reopen) * https://jira.automotivelinux.org/browse/SPEC-2629 (reopen) * 8.0.3 pending * Update 12/10: * 8.0.3 report: https://lists.automotivelinux.org/g/agl-dev-community/message/7869 * 8.99.2 aka RC2 is available at staging location * https://download.automotivelinux.org/AGL/upload/ci/icefish/8.99.2/ * Fuego / Lava support * Trend to execute tests on target by cmdline 'ftc' * Update 09/17: * Volunteers to add to ftc AGL CI ? * qa-testdefinitions are here: https://git.automotivelinux.org/src/qa-testdefinitions/ * Update 11/14: * Reworking https://git.lavasoftware.org/lava/lava/merge_requests/330 * Update 11/26: * New revision submitted upstream, pending review * Update 12/10: * waiting for upstream review * Corentin: * Update 11/26: * h3+kf deployed, m3 standalone deployed * Fix for busybox ip & can * check bluetooth m3 board (with usb adaptor) * check kvm for qemu jobs * Update 12/10 * bluetooth is now conditional to the 'bluetooth' device tag in lava * fixed can test * work rpi4 for ci lab: problem is that there is not network support in u-boot for the pi4 atm. * no netboot right now (trying option with kexec) * Application tests , build multiple wgts * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * Todo: run on qemu targets (new qa-testdefinition) * Update 08/20: * Update 09/03: * no update on build side of wgt * Update 09/17: * no update * Update 10/15: * Waiting for SPEC-2049 * autobuild/linux/autobuild package DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-test DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-debug DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-coverage DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-all DEST=$(pkg)/OUTPUT/ * will go for multiple build targets * Update 11/14: * No update * Update 11/26: * Fred: working on cmake module update, mods in sandbox (https://git.automotivelinux.org/src/cmake-apps-module/commit/?h=sandbox/DDTLK/autobuild&id=0c42794f11bc2d6a0d2d14b3a57f5222a3418b6a) * Move to master and test on app builds * Update 12/10: * migration of apps ? * no update * AGL license scanning (SPDX/fossology) * Update 10/15: * no update * http://95.216.96.43:8081/repo/?mod=browse is for qemux86-64 AGL HH v8.0.0 * Update 11/14: * JIRA for fossology instance: https://jira.automotivelinux.org/browse/SPEC-2957 * Need meta-spdxscanner to be able to create and upload to subfolder * https://jira.automotivelinux.org/browse/SPEC-2961 * Update 11/26: * Instance being setup * Waiting for feature in meta-spdxscanner to upload to subfolder * Update 12/10: * SC approved server. * setting up fossology 3.6 (from docker) * Li: wgt tests only work in guppy and not in halibut * https://jira.automotivelinux.org/browse/SPEC-2795 * Will do app tests on guppy until this is fixed * Update 10/15: * will retest on master * Update 11/14: * waiting for update from Fred on https://jira.automotivelinux.org/browse/SPEC-2795 * Update 11/26: * Fred will cherry-pick the needed changes to * Update 12/10: * No update, added Jose. * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * Update 10/15: * WIP * Update 11/14: * Update 11/26: * postponed after test wgt * Discussion on https://jira.automotivelinux.org/browse/SPEC-2500 * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click * Update 12/10: * Please comment on https://jira.automotivelinux.org/browse/SPEC-3020 * CI Setup document needed - https://jira.automotivelinux.org/browse/SPEC-2980 * Update 12/10: * Files collected. Markdown changes pending. New: * Next call: * Jan 21st 2020 (due to CES) - a good time to everyone through the end of the year !! -------- ==== Meeting 2019/11/26 ==== === Attendees === Jan-Simon, Corentin, Li, Takada-san, Zhou, Ronan, Fred, Kusakabe-san, Juhun ===Topics=== * Fujitsu test report update (Fujitsu) * Update 11/26: * Bugs filed after testing 8.99.1 / RC1 * https://jira.automotivelinux.org/browse/SPEC-2986 * https://jira.automotivelinux.org/browse/SPEC-2987 * https://jira.automotivelinux.org/browse/SPEC-2712 (reopen) * https://jira.automotivelinux.org/browse/SPEC-2630 (reopen) * https://jira.automotivelinux.org/browse/SPEC-2629 (reopen) * 8.0.3 pending * Fuego / Lava support * Trend to execute tests on target by cmdline 'ftc' * Update 09/17: * Volunteers to add to ftc AGL CI ? * qa-testdefinitions are here: https://git.automotivelinux.org/src/qa-testdefinitions/ * Update 11/14: * Reworking https://git.lavasoftware.org/lava/lava/merge_requests/330 * Update 11/26: * New revision submitted upstream, pending review * Corentin: * Update 11/14: * lava-docker 2019-09 deployed * changes for h3 in gerrit * replace qemux86-64 with combined intel-corei7-64 (with qemu output) and move intel build up * insert pi4 * Update 11/26: * h3+kf deployed, m3 standalone deployed * Fix for busybox ip & can * check bluetooth m3 board (with usb adaptor) * check kvm for qemu jobs * Application tests , build multiple wgts * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * Todo: run on qemu targets (new qa-testdefinition) * Update 08/20: * Update 09/03: * no update on build side of wgt * Update 09/17: * no update * Update 10/15: * Waiting for SPEC-2049 * autobuild/linux/autobuild package DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-test DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-debug DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-coverage DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-all DEST=$(pkg)/OUTPUT/ * will go for multiple build targets * Update 11/14: * No update * Update 11/26: * Fred: working on cmake module update, mods in sandbox (https://git.automotivelinux.org/src/cmake-apps-module/commit/?h=sandbox/DDTLK/autobuild&id=0c42794f11bc2d6a0d2d14b3a57f5222a3418b6a) * Move to master and test on app builds * * AGL license scanning (SPDX/fossology) * Update 10/15: * no update * http://95.216.96.43:8081/repo/?mod=browse is for qemux86-64 AGL HH v8.0.0 * Update 11/14: * JIRA for fossology instance: https://jira.automotivelinux.org/browse/SPEC-2957 * Need meta-spdxscanner to be able to create and upload to subfolder * https://jira.automotivelinux.org/browse/SPEC-2961 * Update 11/26: * Instance being setup * Waiting for feature in meta-spdxscanner to upload to subfolder * Li: wgt tests only work in guppy and not in halibut * https://jira.automotivelinux.org/browse/SPEC-2795 * Will do app tests on guppy until this is fixed * Update 10/15: * will retest on master * Update 11/14: * waiting for update from Fred on https://jira.automotivelinux.org/browse/SPEC-2795 * Update 11/26: * Fred will cherry-pick the needed changes to * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * Update 10/15: * WIP * Update 11/14: * Update 11/26: * postponed after test wgt * Discussion on https://jira.automotivelinux.org/browse/SPEC-2500 * Discussion on email reports * example: https://lists.automotivelinux.org/g/agl-test-reports/message/258 * add doc page summarizing existing presentations SPEC-2980 * Follow-up with Khouloud on testreport email * Add link for failures back to lava (best: line of test) * Add link for failures back to jenkins job * tldr: find right section of log with one click New: * Kusakabe-san: * Request for container setup for lava -> https://github.com/kernelci/lava-docker/commit/4e07f8361ddc5677b35b94170b6b1e61b2543b01 * Request for container setup for fuego -> container exists * Setup document needed (re https://jira.automotivelinux.org/browse/SPEC-2980) Next call: * Tuesday Dec 10th ?? (time might change due to F2F !!) -------- ==== Meeting 2019/11/14 ==== === Attendees === Jan-Simon, Li, Stephane, Corentin ===Topics=== * Fujitsu test report update (Fujitsu) * Update 11/14: * No new release. Will test 8.0.3/Icefish rc1 * Fuego / Lava support * Trend to execute tests on target by cmdline 'ftc' * Update 09/17: * Volunteers to add to ftc AGL CI ? * qa-testdefinitions are here: https://git.automotivelinux.org/src/qa-testdefinitions/ * Update 11/14: * Reworking https://git.lavasoftware.org/lava/lava/merge_requests/330 * Corentin: * Update 11/14: * lava-docker 2019-09 deployed * changes for h3 in gerrit * replace qemux86-64 with combined intel-corei7-64 (with qemu output) and move intel build up * insert pi4 * Application tests , build multiple wgts * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * Todo: run on qemu targets (new qa-testdefinition) * Update 08/20: * Update 09/03: * no update on build side of wgt * Update 09/17: * no update * Update 10/15: * Waiting for SPEC-2049 * autobuild/linux/autobuild package DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-test DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-debug DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-coverage DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-all DEST=$(pkg)/OUTPUT/ * will go for multiple build targets * Update 11/14: * No update * AGL license scanning (SPDX/fossology) * Update 10/15: * no update * http://95.216.96.43:8081/repo/?mod=browse is for qemux86-64 AGL HH v8.0.0 * Update 11/14: * JIRA for fossology instance: https://jira.automotivelinux.org/browse/SPEC-2957 * Need meta-spdxscanner to be able to create and upload to subfolder * https://jira.automotivelinux.org/browse/SPEC-2961 * Li: wgt tests only work in guppy and not in halibut * https://jira.automotivelinux.org/browse/SPEC-2795 * Will do app tests on guppy until this is fixed * Update 10/15: * will retest on master * Update 11/14: * waiting for update from Fred on https://jira.automotivelinux.org/browse/SPEC-2795 * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * Update 10/15: * WIP * Update 11/14: * Update to lava-docker 2019-09 * we'll use https://github.com/kernelci/lava-docker/commit/4e07f8361ddc5677b35b94170b6b1e61b2543b01 as a base * Discussion on https://jira.automotivelinux.org/browse/SPEC-2500 New: * Next call: * Tuesday Nov 26th -------- ==== Meeting 2019/10/15 ==== === Attendees === Jan-Simon, Frederic, Ronan, Stephane ===Topics=== * Fujitsu test report update (Fujitsu) * Update 09/17: * No new regressions * mediaplayer addressed by 22428, 22409 * note upsquare behaves differently, might have other issue * radio -> 22382, 22383 * filed can speed: https://jira.automotivelinux.org/browse/SPEC-2811 * 8.0.1 Fujitsu Report links: * {{:eg-ciat:agl_regression_test_hh8.0.1_m3ulcb_kingfisher_m06_20190912.xls|m3+kf}} * {{:eg-ciat:agl_regression_test_hh8.0.1_intel_up2_20190910.xls|up2}} * {{:eg-ciat:agl_regression_test_hh8.0.1_rpi3b_20190911.xls|rpi3}} * {{:eg-ciat:agl_regression_test_hh8.0.1_dragonboard_410c_20190910.xls|db-410c}} * Update 10/15: * no update * Fuego / Lava support * Trend to execute tests on target by cmdline 'ftc' * Update 09/17: * Volunteers to add to ftc AGL CI ? * qa-testdefinitions are here: https://git.automotivelinux.org/src/qa-testdefinitions/ * Lava lab @ IOT.bzh / Frederic: * Update 09/17: * On hold 1m * Update 10/15: * Use 2019-09 of lava-docker: * https://github.com/kernelci/lava-docker/commit/4e07f8361ddc5677b35b94170b6b1e61b2543b01 * Application lifecycle test (Corentin) * Todo: pass/fail criteria for jenkins and/or +/-1 to gerrit. * Update 09/17: * "${BUILD_URL}/console" ?? * https://build.automotivelinux.org/job/ci-platform-meta-agl-demo-verify-CIBT-raspberrypi3/627/injectedEnvVars/ * Update 10/15: * * Application tests * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * Todo: run on qemu targets (new qa-testdefinition) * Update 08/20: * Update 09/03: * no update on build side of wgt * Update 09/17: * no update * Update 10/15: * Waiting for SPEC-2049 * autobuild/linux/autobuild package DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-test DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-debug DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-coverage DEST=$(pkg)/OUTPUT/ * autobuild/linux/autobuild package-all DEST=$(pkg)/OUTPUT/ * will go for multiple build targets * AGL license scanning (SPDX/fossology) * Update 10/15: * no update * http://95.216.96.43:8081/repo/?mod=browse is for qemux86-64 AGL HH v8.0.0 * pi4 works with gfx * Li: wgt tests only work in guppy and not in halibut * https://jira.automotivelinux.org/browse/SPEC-2795 * Will do app tests on guppy until this is fixed * Update 10/15: * will retest on master * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 * Update 10/15: * WIP New: * Update to lava-docker 2019-09 * we'll use https://github.com/kernelci/lava-docker/commit/4e07f8361ddc5677b35b94170b6b1e61b2543b01 as a base * Discussion on https://jira.automotivelinux.org/browse/SPEC-2500 -------- ==== Meeting 2019/09/17 ==== === Attendees === Jan-Simon, Fred, Corentin, Ronan, Li, Walt, Zhou, Kusakabe-san ===Topics=== * Fujitsu test report update (Fujitsu) * Update 09/17: * No new regressions * mediaplayer addressed by 22428, 22409 * note upsquare behaves differently, might have other issue * radio -> 22382, 22383 * filed can speed: https://jira.automotivelinux.org/browse/SPEC-2811 * 8.0.1 Fujitsu Report links: * {{:eg-ciat:agl_regression_test_hh8.0.1_m3ulcb_kingfisher_m06_20190912.xls|m3+kf}} * {{:eg-ciat:agl_regression_test_hh8.0.1_intel_up2_20190910.xls|up2}} * {{:eg-ciat:agl_regression_test_hh8.0.1_rpi3b_20190911.xls|rpi3}} * {{:eg-ciat:agl_regression_test_hh8.0.1_dragonboard_410c_20190910.xls|db-410c}} * Fuego / Lava support * Trend to execute tests on target by cmdline 'ftc' * Update 09/17: * Volunteers to add to ftc AGL CI ? * qa-testdefinitions are here: https://git.automotivelinux.org/src/qa-testdefinitions/ * Lava lab @ IOT.bzh / Frederic: * Update 09/17: * On hold 1m * Application lifecycle test (Corentin) * Todo: pass/fail criteria for jenkins and/or +/-1 to gerrit. * Update 09/17: * Patch in, follow-up in gerrit * Todo: put link to machine specific jenkins log in * https://git.automotivelinux.org/ci-management/tree/jjb/common/include-agl-lava-jobs-submit.sh#n139 * https://git.automotivelinux.org/ci-management/tree/jjb/common/include-agl-lava-jobs-submit.sh#n142 * e.g. link to https://build.automotivelinux.org/job/ci-platform-meta-agl-demo-verify-CIBT-raspberrypi3/627/console * https://build.automotivelinux.org/job/ci-platform-meta-agl-demo-verify-CIBT-raspberrypi3/627/console * "${BUILD_URL}/console" ?? * https://build.automotivelinux.org/job/ci-platform-meta-agl-demo-verify-CIBT-raspberrypi3/627/injectedEnvVars/ * Application tests * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * Todo: run on qemu targets (new qa-testdefinition) * Update 08/20: * Update 09/03: * no update on build side of wgt * Update 09/17: * no update * AGL license scanning (SPDX/fossology) * Meeting with Michael Jäger (fossology): we need to start using the rest api in fossdriver (instead of scraping the web-page) * Follow-up discussions on these alternatives * https://github.com/heremaps/oss-review-toolkit * https://github.com/endocode * Update 08/20: * To execute only spdx tasks (and its dependencies like unpack/patch) execute: * bitbake --runall=spdx agl-demo-platform * http://95.216.96.43:8081/repo/?mod=browse (default credentials) * AI: Kusakabe-san and team to go through it * Update 09/03: * Docker container was down - re-populate (Jan-Simon) * Update 09/17: * Trouble re-creating working environment (fossology 3.3.0 + meta-spdxscanner New: * Li: rpi4 support -> with gfx planned for Icefish (patches welcome) * Stephane: 64bit kernel on pi's ? No, due to gfx stack status and memory footprint * Zhou: apps-agl-cluster-demo-gauges-owners - done. * Li: wgt tests only work in guppy and not in halibut * https://jira.automotivelinux.org/browse/SPEC-2795 * Will do app tests on guppy until this is fixed * HTML5 apps, CI integration with ./autobuild/autobuild script * https://jira.automotivelinux.org/browse/SPEC-2773 -------- ==== Meeting 2019/09/03 ==== === Attendees === Jan-Simon, Frederic, Li, Zhou, Kurokawa-san ===Topics=== * Fujitsu test report update (Fujitsu) * Update 09/03: * No update * Fuego / Lava support * Trend to execute tests on target by cmdline 'ftc' * Update 09/03: * No update * Lava lab @ IOT.bzh / Frederic: * Update 09/03: * No update * qemu worker if possible * Application lifecycle test (Corentin) * Todo: pass/fail criteria for jenkins and/or +/-1 to gerrit. * Working out pass/fail/skip for CIB / CIBT * how can app developers define pass/fail/skip * how can testdefinitions interpret * jenkins needs possibility to skip certain cases (and not fail) * -test wgt's can have multiple invocations (e.g. representing different test plans) * Update 08/20: * Discussed https://jira.automotivelinux.org/browse/SPEC-2703?focusedCommentId=21708&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-21708 * Will go with qa-testdefinition or app repo * Update 09/03: * Patch in gerrit for review * go with "'ci-platform' in the JOB_NAME or JOB_BASE_NAME" * Application tests * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * Todo: run on qemu targets (new qa-testdefinition) * Update 08/20: * Update 09/03: * no update on build side of wgt * AGL license scanning (SPDX/fossology) * Meeting with Michael Jäger (fossology): we need to start using the rest api in fossdriver (instead of scraping the web-page) * Follow-up discussions on these alternatives * https://github.com/heremaps/oss-review-toolkit * https://github.com/endocode * Update 08/20: * To execute only spdx tasks (and its dependencies like unpack/patch) execute: * bitbake --runall=spdx agl-demo-platform * http://95.216.96.43:8081/repo/?mod=browse (default credentials) * AI: Kusakabe-san and team to go through it * Update 09/03: * Docker container was down - re-populate (Jan-Simon) On hold: * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * Update 11/12: Jobs rework to use PRSERV (but will invalidate sstate-cache) * Update 14/05: PRSERV use reverted as it caused issues * AI (jsm): bitbake package-index * AI (jsm): sync-out the rpm feeds for the snapshot builds * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * AI (Tom): is there a prserv (ro) version? - Likely no. New: * ---- ==== Meeting 2019/08/20 ==== === Attendees === Jan-Simon, Kurokawa-san, Frederic, Corentin, Li Xiaoming, Stephane, ===Topics=== * Fujitsu test report update (Fujitsu) * Update 08/20: * No update. * Fuego / Lava support * 06/25 : No update * 07/09 : No update * Trend to execute tests on target by cmdline 'ftc' * Update 08/20: * No update * Lava lab @ IOT.bzh / Frederic: * lava-docker version of master: 19fe02f7a341e302e3ad560fcb4cc872f2201a5c * 07/23 : Still in a box * Update 08/20: * No update * Application lifecycle test (Corentin) * Todo: pass/fail criteria for jenkins and/or +/-1 to gerrit. * Working out pass/fail/skip for CIB / CIBT * how can app developers define pass/fail/skip * how can testdefinitions interpret * jenkins needs possibility to skip certain cases (and not fail) * -test wgt's can have multiple invocations (e.g. representing different test plans) * Update 08/20: * Discussed https://jira.automotivelinux.org/browse/SPEC-2703?focusedCommentId=21708&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-21708 * Will go with qa-testdefinition or app repo * Application tests * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * Todo: run on qemu targets (new qa-testdefinition) * Update 08/20: * no update on build side of wgt * AGL license scanning (SPDX/fossology) * Meeting with Michael Jäger (fossology): we need to start using the rest api in fossdriver (instead of scraping the web-page) * Follow-up discussions on these alternatives * https://github.com/heremaps/oss-review-toolkit * https://github.com/endocode * Update 08/20: * To execute only spdx tasks (and its dependencies like unpack/patch) execute: * bitbake --runall=spdx agl-demo-platform * http://95.216.96.43:8081/repo/?mod=browse (default credentials) * AI: Kusakabe-san and team to go through it On hold: * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * Update 11/12: Jobs rework to use PRSERV (but will invalidate sstate-cache) * Update 14/05: PRSERV use reverted as it caused issues * AI (jsm): bitbake package-index * AI (jsm): sync-out the rpm feeds for the snapshot builds * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * AI (Tom): is there a prserv (ro) version? - Likely no. New: * -------- ==== Meeting 2019/07/09 ==== === Attendees === Jan-Simon, Frederic, Ronan, Kurokawa-san, Abe-san, Li-Xiaoming, Kusakabe-san, Stephane ===Topics=== * Fujitsu test report update (Fujitsu) * RC1 - new jira: * https://jira.automotivelinux.org/browse/SPEC-2547 (Wifi status icon) * https://jira.automotivelinux.org/browse/SPEC-2551 (Mediaplayer /& BT) * https://jira.automotivelinux.org/browse/SPEC-2552 (Upsquare / avirt) * https://jira.automotivelinux.org/browse/SPEC-2553 (dragonboard boot) * RC2 - * https://jira.automotivelinux.org/browse/SPEC-2616 (rpi shows no homescreen) * https://jira.automotivelinux.org/browse/SPEC-2552 (Upsquare also rc2) * https://jira.automotivelinux.org/browse/SPEC-2553 (dragonboard boot also rc2) * m3 testing WIP * RC3 - * Starting tomorrow. * Fuego / Lava support * 06/25 : No update * 07/09 : No update * Trend to execute tests on target by cmdline 'ftc' * Lava lab @ IOT.bzh / Frederic: * lava-docker version of master: 19fe02f7a341e302e3ad560fcb4cc872f2201a5c * 06/25 : No update * 07/09 : Still in a box * 07/23 : Still in a box * Application lifecycle test (Corentin) * Update 06/25: * qemuarm / qemuarm64 added. * Todo: pass/fail criteria for jenkins and/or +/-1 to gerrit. * Application tests * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * Todo: run on qemu targets (new qa-testdefinition) * Working out pass/fail/skip for CIB / CIBT * how can app developers define pass/fail/skip * how can testdefinitions interpret * jenkins needs possibility to skip certain cases (and not fail) * -test wgt's can have multiple invocations (e.g. representing different test plans) * AGL license scanning (SPDX/fossology) * -native recipe for tools to scan w/o fossology server (dosocs-native, ) * alternative is to use fossology server and include a fossup-native (fossdriver? XYZ(fork of fossup) ) in the build and provide the necessary variables. * Goal is to enable spdx scanning for the release build. * Update 07/09: Jan-Simon tried all three options (fossdriver, dosocs-host, dosocs) * need to file further bugs to be fixed * fossology version for fossdriver (not latest container): use 3.3.0 * reproduce: * fossdriver: just add agl-fossdriver feature * dosocs: add agl-fossdriver feature and change local.conf from * INHERIT += "fossdriver" to * INHERIT += "dosocs" * if possible sample run for ALS to compare with existing SPDXLite data (for Kusakabe-san) * Update 07/23: * - POC using https://github.com/dl9pf/meta-spdxscanner/tree/jsm-thud and fossology container version 3.3 * Quick resetup: * a) server host with: docker run -d -p 8081:80 fossology/fossology:3.3.0 #Needs to be 3.3!!! * b) compile host with: * b1) fossdriver installed on build host https://github.com/fossology/fossdriver * b2) ~/.fossdriverrc * b3) conf/local.conf: * INHERIT += "fossdriver-host" * SPDX_DEPLOY_DIR = "${TOPDIR}/spdx/" * - POC up here: http://95.216.96.43:8081/repo/?mod=browse * Discussed during F2F in Yokohama * TLDR: We will run it for the HH 8.0.0 release * will merely record the parsed data and write the manifest as-is * clearing/postprocessing done by Fujitsu * Meeting with Michael Jäger (fossology): we need to start using the rest api in fossdriver (instead of scraping the web-page) * Follow-up discussions on these alternatives * https://github.com/heremaps/oss-review-toolkit * https://github.com/endocode On hold: * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * Update 11/12: Jobs rework to use PRSERV (but will invalidate sstate-cache) * Update 14/05: PRSERV use reverted as it caused issues * AI (jsm): bitbake package-index * AI (jsm): sync-out the rpm feeds for the snapshot builds * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * AI (Tom): is there a prserv (ro) version? - Likely no. New: * -------- ==== Meeting 2019/07/09 ==== === Attendees === Jan-Simon, Frederic, Stephane, Li Xiaoming, Zhou Mingying, Ronan, Kusakabe-san ===Topics=== * Fujitsu test report update (Fujitsu) * RC1 - new jira: * https://jira.automotivelinux.org/browse/SPEC-2547 (Wifi status icon) * https://jira.automotivelinux.org/browse/SPEC-2551 (Mediaplayer /& BT) * https://jira.automotivelinux.org/browse/SPEC-2552 (Upsquare / avirt) * https://jira.automotivelinux.org/browse/SPEC-2553 (dragonboard boot) * RC2 - * https://jira.automotivelinux.org/browse/SPEC-2616 (rpi shows no homescreen) * https://jira.automotivelinux.org/browse/SPEC-2552 (Upsquare also rc2) * https://jira.automotivelinux.org/browse/SPEC-2553 (dragonboard boot also rc2) * m3 testing WIP * Fuego / Lava support * 06/25 : No update * 07/09 : No update * Lava lab @ IOT.bzh / Frederic: * lava-docker version of master: 19fe02f7a341e302e3ad560fcb4cc872f2201a5c * 06/25 : No update * 07/09 : Still in a box * Application lifecycle test (Corentin) * Update 06/25: * qemuarm / qemuarm64 added. * Todo: pass/fail criteria for jenkins and/or +/-1 to gerrit. * Application tests * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * Todo: run on qemu targets (new qa-testdefinition) * Working out pass/fail/skip for CIB / CIBT * how can app developers define pass/fail/skip * how can testdefinitions interpret * jenkins needs possibility to skip certain cases (and not fail) * -test wgt's can have multiple invocations (e.g. representing different test plans) * AGL license scanning (SPDX/fossology) * -native recipe for tools to scan w/o fossology server (dosocs-native, ) * alternative is to use fossology server and include a fossup-native (fossdriver? XYZ(fork of fossup) ) in the build and provide the necessary variables. * Goal is to enable spdx scanning for the release build. * Update 07/09: Jan-Simon tried all three options (fossdriver, dosocs-host, dosocs) * need to file further bugs to be fixed * fossology version for fossdriver (not latest container): use 3.3.0 * reproduce: * fossdriver: just add agl-fossdriver feature * dosocs: add agl-fossdriver feature and change local.conf from * INHERIT += "fossdriver" to * INHERIT += "dosocs" * if possible sample run for ALS to compare with existing SPDXLite data (for Kusakabe-san) On hold: * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * Update 11/12: Jobs rework to use PRSERV (but will invalidate sstate-cache) * Update 14/05: PRSERV use reverted as it caused issues * AI (jsm): bitbake package-index * AI (jsm): sync-out the rpm feeds for the snapshot builds * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * AI (Tom): is there a prserv (ro) version? - Likely no. New: * -------- ==== Meeting 2019/06/25 ==== === Attendees === Jan-Simon, Li, Ronan, Kusakabe-san, Stephane, Corentin ===Topics=== * Fujitsu test report update (Fujitsu) * RC1 - new jira: * https://jira.automotivelinux.org/browse/SPEC-2547 (Wifi status icon) * https://jira.automotivelinux.org/browse/SPEC-2551 (Mediaplayer /& BT) * https://jira.automotivelinux.org/browse/SPEC-2552 (Upsquare / avirt) * https://jira.automotivelinux.org/browse/SPEC-2553 (dragonboard boot) * Fuego / Lava support * 06/25 : No update * Lava lab @ IOT.bzh / Frederic: * lava-docker version of master: 19fe02f7a341e302e3ad560fcb4cc872f2201a5c * 06/25 : No update * Application lifecycle test (Corentin) * Update 06/25: * qemuarm / qemuarm64 added. * Todo: pass/fail criteria for jenkins and/or +/-1 to gerrit. * Application tests * Edi has problems reproducing the test wgt's * Test wgt's for apps submitted * Todo: build test wgt in ci (Jan-Simon) * Todo: extend releng/qa-testdefinitions to be able to download a 2nd wgt file (e.g. homescreen-test.wgt) * Todo: run on qemu targets (new qa-testdefinition) * Working out pass/fail/skip for CIB / CIBT * how can app developers define pass/fail/skip * how can testdefinitions interpret * jenkins needs possibility to skip certain cases (and not fail) * -test wgt's can have multiple invocations (e.g. representing different test plans) * AGL license scanning (SPDX/fossology) * -native recipe for tools to scan w/o fossology server (dosocs-native, ) * alternative is to use fossology server and include a fossup-native (fossdriver? XYZ(fork of fossup) ) in the build and provide the necessary variables. * Goal is to enable spdx scanning for the release build. On hold: * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * Update 11/12: Jobs rework to use PRSERV (but will invalidate sstate-cache) * Update 14/05: PRSERV use reverted as it caused issues * AI (jsm): bitbake package-index * AI (jsm): sync-out the rpm feeds for the snapshot builds * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * AI (Tom): is there a prserv (ro) version? - Likely no. New: * ==== Meeting 2019/05/28 ==== ===Attendees=== Jan-Simon, Frederic, Kusakabe-san, Li Xiaoming, Stephane ===Topics=== * Fujitsu test report update (Fujitsu) * Update 05/28: * tested mediaplayer with USB soundcard on RPI3b resulted in crash of mediaplayer * check sound device path in hal file ? /usr/libexec/agl/4a-hal/etc * against ls /dev/snd/by-id/ * does 4a-play work ? -> yes * 2nd: what are the messages in syslog (crash) -> please file a jira and attach. * Stephane reported some possible runtime issues on m4 (might be related to avirt, snd-aloop seems to work) * Fuego / Lava support * Being reworked and re-submitted to lava. * Update 05/14: No update * Update 05/28: Change submitted. * Application lifecycle test * Update 05/14: * Need to fix jobs with branches != master (e.g. http://lava.automotivelinux.org/scheduler/job/3496 ). Corentin will try to add branch. * Todo: pass/fail criteria for jenkins and/or +/-1 to gerrit. * Infra update: * Update 05/14: No update. * Update 05/28: No update. * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * Update 11/12: Jobs rework to use PRSERV (but will invalidate sstate-cache) * Update 14/05: PRSERV use reverted as it caused issues * AI (jsm): bitbake package-index * AI (jsm): sync-out the rpm feeds for the snapshot builds * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * AI (Tom): is there a prserv (ro) version? - Likely no. * Update 05/14: no update * Update 05/28: no update * Frederic: * lava-docker version of master: 19fe02f7a341e302e3ad560fcb4cc872f2201a5c * Update 05/14: * Update done. Cleaned-up master and restarting docker. * https://jira.automotivelinux.org/browse/SPEC-2453 * https://jira.automotivelinux.org/browse/SPEC-2454 * Lab will move to new office mid-june. * Application tests * Edi has problems reproducing the test wgt's * will file a jira ticket (@Romain or @Frederic) * Update 05/14: * Need to decide if we run tests with real or virtualized bt hw. * BT connect is there, but inactive atm until ^^^ . * -> devcall or SAT. * Update 05/28: no update New: * Li: * AGL license scanning (SPDX/fossology) - report * Stephane: * https://github.com/dl9pf/meta-spdxscanner possibly adding a dosocs-native recipe and config fragment so we can just turn it on by feature. ==== Meeting 2019/05/14 ==== ===Attendees=== Jan-Simon, Edi, Frederic, Khouloud, Stephane, Corentin, Tom ===Topics=== * Fujitsu test report update (Fujitsu) * Update 04/30: golden week * Update 05/14: No update * Fuego / Lava support * Being reworked and re-submitted to lava. * Update 04/30: golden week * Update 05/14: No update * Application lifecycle test * Update 04/16: https://gerrit.automotivelinux.org/gerrit/#/c/20955/ done. * Update 04/30: Jobs being submitted, hunting bugs * releng-scripts/utils/create-jobs.py --machine qemux86-64 --build-type snapshot --branch master --version latest --app-changeid 21120 --app-patchset 2 --name AGL-gerrit-21120-2 --build-version AGL-gerrit-21120-2 --callback-from lab-agl-core * Need to change snapshot to use -crosssdk * '--build-type' need better --help and doc * @Corentin/Khouloud, please take a look. Tnx. * Update 05/14: * Need to fix jobs with branches != master (e.g. http://lava.automotivelinux.org/scheduler/job/3496 ). Corentin will try to add branch. * Todo: pass/fail criteria for jenkins and/or +/-1 to gerrit. * Infra update: * Update: 04/16 - no news * Update 04/30: moving release files to mirrors.kernel.org * Update 05/14: No update. * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * Update 11/12: Jobs rework to use PRSERV (but will invalidate sstate-cache) * Update 14/05: PRSERV use reverted as it caused issues * AI (jsm): bitbake package-index * AI (jsm): sync-out the rpm feeds for the snapshot builds * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * AI (Tom): is there a prserv (ro) version? - Likely no. * Update 2019-04-02: busy with thud uprev * Update 04/16: no update * Update 04/30: no update * Update 05/14: no update * Frederic: * lava-docker version of master: 19fe02f7a341e302e3ad560fcb4cc872f2201a5c * Update 04/02: board up, but we need the next lava-docker version to have the device-type supported * Update 04/30: * branch AGL master uses https://github.com/montjoie/lava-docker/commits/agl-2019.03 * Doc is https://github.com/montjoie/lava-docker/commit/2fb6759a52e2e7ba267d0c0b93c1b32d4bd98dcb for the dhcp server setup for pxe (adapt IPs as necessary). * Update 05/14: * Update done. Cleaned-up master and restarting docker. * Talk about layer structure and moving components into the core * proposal update to thud first, then take a look at all components (add/move/remove) * meta-agl-devel/* * meta-oic * Update 04/16: * meta-oic removed already from master * need to talk about in SAT: * meta-agl-devel/meta-audio-4a-framework * meta-agl-devel/meta-speech-framework * meta-agl-devel/meta-hmi-framework ** * Update 04/30: * Next SAT / F2F (slot on wed) * Update 05/14: * Talked during F2F * https://docs.google.com/presentation/d/1b8VcJkymovpY6bnwUIsfERtGwfm_I62MkfxMt6cJJsU/edit?usp=sharing * Folder structure for AGL repo clone: * Sort projects into subfolder ?: * meta-agl/ * meta-agl-*/ * community-bsp/ * reference-bsp/ * own-project/ * upstream/ * Update 04/30: * Next SAT / F2F (slot on wed) * Update 05/14: * Talked during F2F * https://docs.google.com/presentation/d/1b8VcJkymovpY6bnwUIsfERtGwfm_I62MkfxMt6cJJsU/edit?usp=sharing * Application tests * Edi has problems reproducing the test wgt's * will file a jira ticket (@Romain or @Frederic) * Update 04/16: helloworld test now runs (SPEC-2314) * Update 04/16: agl-service-gps * Update 04/30: testing various services now * bluetooth-service, weather-service, mediaplayer-service * some issues with afm-test (crash?? on 1 out of 3 tries) * please file a jira and assign to Romain * Need to investigate how to virtualize bt device, wifi device * https://github.com/martinpitt/umockdev * Wrt Hardware: check if HW is up and we get a valid response from the hardware * Wrt services/api's: check api's agains virtual or mock'd device. * Update 05/14: * Need to decide if we run tests with real or virtualized bt hw. * BT connect is there, but inactive atm until ^^^ . * -> devcall or SAT. New: * ==== Meeting 2019/04/30 ==== ===Attendees=== Jan-Simon, Gmike, Stephane, Corentin, Khouloud ===Topics=== * Fujitsu test report update (Fujitsu) * Update 04/30: golden week * Fuego / Lava support * Being reworked and re-submitted to lava. * Update 04/30: golden week * Application lifecycle test * Update 04/16: https://gerrit.automotivelinux.org/gerrit/#/c/20955/ done. * Update 04/30: Jobs being submitted, hunting bugs * releng-scripts/utils/create-jobs.py --machine qemux86-64 --build-type snapshot --branch master --version latest --app-changeid 21120 --app-patchset 2 --name AGL-gerrit-21120-2 --build-version AGL-gerrit-21120-2 --callback-from lab-agl-core * Need to change snapshot to use -crosssdk * '--build-type' need better --help and doc * @Corentin/Khouloud, please take a look. Tnx. * Infra update: * Update: 04/16 - no news * Update 04/30: moving release files to mirrors.kernel.org * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * PRSERV setup in infra, addition to jobs pending * Update 13/11: prserv setup, need to use in jobs * Update 11/12: Jobs rework to use PRSERV (but will invalidate sstate-cache) * AI (jsm): bitbake package-index * AI (jsm): sync-out the rpm feeds for the snapshot builds * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * AI (Tom): is there a prserv (ro) version? - Likely no. * Update 2019-04-02: busy with thud uprev * Update 04/16: no update * Update 04/30: no update * issue with ptest packages, CI-build should enable agl-ptest * https://gerrit.automotivelinux.org/gerrit/17505 * tbd: selection of tests as some break the network filesystem * take board, run nbdroot and execute ptests one-by-one -> list of good/bad * Update 04/02: cannot reproduce locally - test on all boards in CI with snapshot * Update 04/16: no update * Update 04/30: systemd moved to meson. ptest was removed. Issue closed. * Frederic: * lava-docker version of master: 19fe02f7a341e302e3ad560fcb4cc872f2201a5c * Update 04/02: board up, but we need the next lava-docker version to have the device-type supported * Update 04/30: * branch AGL master uses https://github.com/montjoie/lava-docker/commits/agl-2019.03 * Doc is https://github.com/montjoie/lava-docker/commit/2fb6759a52e2e7ba267d0c0b93c1b32d4bd98dcb for the dhcp server setup for pxe (adapt IPs as necessary). * Lava-docker update: * Update: 04/16: lava-docker release with fix this week, uprev planned for next week * Update 04/30: deployed * Talk about layer structure and moving components into the core * proposal update to thud first, then take a look at all components (add/move/remove) * meta-agl-devel/* * meta-oic * Update 04/16: * meta-oic removed already from master * need to talk about in SAT: * meta-agl-devel/meta-audio-4a-framework * meta-agl-devel/meta-speech-framework * meta-agl-devel/meta-hmi-framework ** * Update 04/30: * Next SAT / F2F (slot on wed) * Folder structure for AGL repo clone: * Sort projects into subfolder ?: * AGL/ * community-bsp/ * reference-bsp/ * own-project/ * upstream/ * Update 04/30: * Next SAT / F2F (slot on wed) * Application tests * Edi has problems reproducing the test wgt's * will file a jira ticket (@Romain or @Frederic) * Update 04/16: helloworld test now runs (SPEC-2314) * Update 04/16: agl-service-gps * Update 04/30: testing various services now * bluetooth-service, weather-service, mediaplayer-service * some issues with afm-test (crash?? on 1 out of 3 tries) * please file a jira and assign to Romain * Need to investigate how to virtualize bt device, wifi device * https://github.com/martinpitt/umockdev * Wrt Hardware: check if HW is up and we get a valid response from the hardware * Wrt services/api's: check api's agains virtual or mock'd device. New: * gerrit and jira page load slow - filed https://jira.automotivelinux.org/browse/SPEC-2370 ==== Meeting 2019/04/16 ==== THIS MEETING WILL STILL USE GOTOMEETING ! ===Attendees=== Jan-Simon, Li, Edi, Stoyan, Corentin, Stephane, Kusakabe-san ===Topics=== * Fujitsu test report update (Fujitsu) * 7.0.1 w/ dragonboard 410c * https://jira.automotivelinux.org/browse/SPEC-2324 * likely kernel bug (-> BSP) * Fuego / Lava support * Being reworked and re-submitted to lava. * Application lifecycle test * Update 04/16: https://gerrit.automotivelinux.org/gerrit/#/c/20955/ done. * Thud upgrade * Status in weekly devcall * CI updates necessary * dtb names changed * fix required in releng. * arm32 SDK name changed armv7ve... -> armv7vet2... * Infra update: * new CI infra is WIP, change in provider, build nodes will be physical machines * Update: 04/16 - no news * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * PRSERV setup in infra, addition to jobs pending * Update 13/11: prserv setup, need to use in jobs * Update 11/12: Jobs rework to use PRSERV (but will invalidate sstate-cache) * AI (jsm): bitbake package-index * AI (jsm): sync-out the rpm feeds for the snapshot builds * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * AI (Tom): is there a prserv (ro) version? - Likely no. * Update 2019-04-02: busy with thud uprev * Update 04/16: no update * issue with ptest packages, CI-build should enable agl-ptest * https://gerrit.automotivelinux.org/gerrit/17505 * tbd: selection of tests as some break the network filesystem * take board, run nbdroot and execute ptests one-by-one -> list of good/bad * Update 02/19 * AI for baylibre to fix systemd test * Removing network sub-tests from systemd ptest (due to netboot environment) * Update 04/02: cannot reproduce locally - test on all boards in CI with snapshot * Update 04/16: no update * Frederic: * lava-docker version of master: 19fe02f7a341e302e3ad560fcb4cc872f2201a5c * Update 04/02: board up, but we need the next lava-docker version to have the device-type supported * Lava-docker update: * already testing and PR issued * issue is pxe+grub+nbdroot (still pending in upstream) * uprev of master and slaves needed for this * schedule during next call * Update: 04/16: lava-docker release with fix this week, uprev planned for next week * Talk about layer structure and moving components into the core * proposal update to thud first, then take a look at all components (add/move/remove) * meta-agl-devel/* * meta-oic * Update 04/16: * meta-oic removed already from master * need to talk about in SAT: * meta-agl-devel/meta-audio-4a-framework * meta-agl-devel/meta-speech-framework * meta-agl-devel/meta-hmi-framework * Folder structure for AGL repo clone: * Sort projects into subfolder ?: * AGL/ * community-bsp/ * reference-bsp/ * own-project/ * upstream/ * Application tests * Edi has problems reproducing the test wgt's * will file a jira ticket (@Romain or @Frederic) * Update 04/16: helloworld test now runs (SPEC-2314) * Update 04/16: agl-service-gps intel-corei7-64:~# afm-test agl-service-gps-test.wgt PASS: agl-service-gps-test@1.0-2c2b36d-dirty started with pid=13949 ERROR: failed Resource temporarily unavailable PASS: agl-service-gps-test@1.0-2c2b36d-dirty killed and removed * Try to build with IMAGE_INSTALL_append = " gps-utils" * If that works we need to add the dependency in the proper place (agl-service-gps if agl-devel is enabled) New: * Switch to zoom !! ==== Meeting 2019/04/02 ==== ===Attendees=== Jan-Simon, Tom, Edi, Stoyan, Li, Ronan, Corentin, Stephane, Frederic, Khouloud ===Topics=== * Fujitsu test report update (Fujitsu) * GG 7.0.1 tested * test reports: {{:eg-ciat:aglgg7_0_1regressiontestreport.zip|testreport xml zipped}} * Issues with USB radio, BT, hvac and dashboard * JIRA's to be filed or extended * Fuego / Lava support * Presentation to be held during AGL AMM * https://aglammjapan2019.sched.com/event/L8VH/enhanced-ciat-lab-for-agl-running-lava-with-fuegos-binary-test-cache-zhong-lu-li-gang-nanjing-fujitsu-nanda-software-technology-co-ltd * Test case submissions to JIRA (review) * https://jira.automotivelinux.org/projects/SPEC?selectedItem=com.thed.zephyr.je%3Azephyr-tests-page#test-cycles-tab * 04/02: Ok so far * Application lifecycle test * Where to add app test: * https://git.automotivelinux.org/ci-management/tree/jjb/ci-apps-verify/ci-apps-jjb.yaml * we need to extend this with CIBT (multijob) similar to https://git.automotivelinux.org/ci-management/tree/jjb/ci-AGL-verify/ci-platform-jjb.yaml#n150 * initially just qemux86-64 / image from https://download.automotivelinux.org/AGL/snapshots/master/latest/qemux86-64/deploy/images/qemux86-64/ for lava job Along the lines of: - multijob: name: CIBT-app-{project} condition: SUCCESSFUL projects: - name: 'ci-app-{project}-verify-CIBT-qemux86-64' current-parameters: true #abort-all-job: true kill-phase-on: FAILURE - multijob: name: 'CIBT-{project}-finish-success' condition: SUCCESSFUL projects: - name: 'ci-app-{project}-verify-CIBT-finish-success' current-parameters: true abort-all-job: true kill-phase-on: FAILURE * Thud upgrade * Status in weekly devcall * master-next branch: * https://jira.automotivelinux.org/browse/SPEC-1837 * https://git.automotivelinux.org/AGL/meta-agl/commit/?h=next&id=cd14a879b91fb74e58c01e31248590ce5351c94e * ''repo grep AGL_FIXME'' * ''find . -name *AGLFIXME*'' * and you shall find the interesting spots to fix. * Update 02/19: See mail on ML and sandbox * Infra update: * new CI infra is WIP, change in provider, build nodes will be physical machines * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * PRSERV setup in infra, addition to jobs pending * Update 13/11: prserv setup, need to use in jobs * Update 11/12: Jobs rework to use PRSERV (but will invalidate sstate-cache) * AI (jsm): bitbake package-index * AI (jsm): sync-out the rpm feeds for the snapshot builds * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * AI (Tom): is there a prserv (ro) version? - Likely no. * UPDATE 2019-01-22: Work resumed * Update 2019-02-19: Busy with GG7.0.0 * Update 2019-04-02: busy with thud uprev * issue with ptest packages, CI-build should enable agl-ptest * https://gerrit.automotivelinux.org/gerrit/17505 * tbd: selection of tests as some break the network filesystem * take board, run nbdroot and execute ptests one-by-one -> list of good/bad * Update 02/19 * AI for baylibre to fix systemd test * Removing network sub-tests from systemd ptest (due to netboot environment) * Update 04/02: cannot reproduce locally - test on all boards in CI with snapshot * Working on app test for ci-apps-* (https://jira.automotivelinux.org/browse/SPEC-1850) * WIP, close to be complete * Frederic: * lava-docker version of master: 19fe02f7a341e302e3ad560fcb4cc872f2201a5c * Update 04/02: board up, but we need the next lava-docker version to have the device-type supported * Lava-docker update: * already testing and PR issued * issue is pxe+grub+nbdroot (still pending in upstream) * uprev of master and slaves needed for this * schedule during next call * Talk about layer structure and moving components into the core * proposal update to thud first, then take a look at all components (add/move/remove) * meta-agl-devel/* * meta-oic * 02/19: defer after update * 04/02: " * Ronan: RCAR-BSP 3.15.0 is ready for review in gerrit * 04/02: done NEW: * Application tests * Edi has problems reproducing the test wgt's * will file a jira ticket (@Romain or @Frederic) * === Next meeting === Next meeting in 2 weeks. ATTENTION: we'll switch to zoom ! Watch-out for an updated calendar-invite. ==== Meeting 2019/02/19 ==== ===Attendees=== Jan-Simon, Frederic, Kurokawa-san, Li Xiaoming, Ronan, Tom, Khouloud, Corentin ===Topics=== * Fujitsu test report update * Guppy 7.0.0: found Dashboard and HVAC crashing during test execution. -> JIRA (assign to Leon) * Test case submissions to JIRA * Test cases can be defined in Jira now. * Example: https://jira.automotivelinux.org/browse/SPEC-2132 * Todo: document tests executed * Update 2/19: Sample entries for 'Zephyr' [[https://jira.automotivelinux.org/projects/SPEC?selectedItem=com.thed.zephyr.je%3Azephyr-tests-page#test-cycles-tab | here] * Fuego / Lava support * Patch for LAVA pending review https://git.lavasoftware.org/lava/lava/merge_requests/330 * Presentation to be held during AGL AMM * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * PRSERV setup in infra, addition to jobs pending * Update 13/11: prserv setup, need to use in jobs * Update 11/12: Jobs rework to use PRSERV (but will invalidate sstate-cache) * AI (jsm): bitbake package-index * AI (jsm): sync-out the rpm feeds for the snapshot builds * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * AI (Tom): is there a prserv (ro) version? - Likely no. * UPDATE 2019-01-22: Work resumed * Update 2019-02-19: Busy with GG7.0.0 * master-next branch: * https://jira.automotivelinux.org/browse/SPEC-1837 * https://git.automotivelinux.org/AGL/meta-agl/commit/?h=next&id=cd14a879b91fb74e58c01e31248590ce5351c94e * ''repo grep AGL_FIXME'' * ''find . -name *AGLFIXME*'' * and you shall find the interesting spots to fix. * Update 02/19: See mail on ML and sandbox * issue with ptest packages, CI-build should enable agl-ptest * https://gerrit.automotivelinux.org/gerrit/17505 * tbd: selection of tests as some break the network filesystem * take board, run nbdroot and execute ptests one-by-one -> list of good/bad * Update 02/19 * AI for baylibre to fix systemd test * Removing network sub-tests from systemd ptest (due to netboot environment) * 02/19: deferred * Working on app test for ci-apps-* (https://jira.automotivelinux.org/browse/SPEC-1850) * WIP, close to be complete * SPEC-1902, Stephane will do the reference features, Jan-Simon the CI features. * Update 02/19 * Some more updates done * Frederic: * lava-docker version of master: 19fe02f7a341e302e3ad560fcb4cc872f2201a5c * working on 64bit rpi3 filesystem * still issues on rpi3 * -> proposal: use x86 board or machine * Jira-Zephyr vs Kiwi vs ??? * Kiwi: * + attach artifacts to the test (stored in kiwi) * repo manifest -r * + formatted output -> pdf * Zephyr: * create dummy set , plan and ad-hoc and planned execution * duplicate build+test+core UI plans from Konsulko to compare (Tom) * No update: * * Talk about layer structure and moving components into the core * proposal update to thud first, then take a look at all components (add/move/remove) * meta-agl-devel/* * meta-oic * 02/19: defer after update NEW: * Ronan: RCAR-BSP 3.15.0 is ready for review in gerrit * Gerrit: https://gerrit.automotivelinux.org/gerrit/#/q/owner:ronan.lemartret%2540iot.bzh+status:open * Sandbox: https://git.automotivelinux.org/AGL/meta-renesas-rcar-gen3/log/?h=sandbox/ronan/rocko_bsp_v3.15.0 * Plan: * for current master and to be cherry-picked to guppy and flounder (3.15.0+rocko) * master-next (thud) will also use 3.15.0+thud for now (push mar for BSP and apr for KF) ==== Meeting 2019/02/04 ==== ===Attendees=== Jan-Simon, Walt, Tom, Frederic, Li, Ronan, Stephane ===Topics=== * Test case submissions to JIRA * Test cases can be defined in Jira now. * Example: https://jira.automotivelinux.org/browse/SPEC-2132 * Todo: document tests executed * AI: Fujitsu & Konsulko - document test cases in JIRA * Fuego / Lava support * Patch for LAVA pending review https://git.lavasoftware.org/lava/lava/merge_requests/330 * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * PRSERV setup in infra, addition to jobs pending * Update 13/11: prserv setup, need to use in jobs * Update 11/12: Jobs rework to use PRSERV (but will invalidate sstate-cache) * Update 11/12: postponed after CES. * Update 02/05: work started -> guppy final * AI (jsm): bitbake package-index * Update 11/12: postponed after CES. * Update 02/05: work started -> guppy final * AI (jsm): sync-out the rpm feeds for the snapshot builds * Update 11/12: more space on download server requested * Update 02/05: work started -> guppy final * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * Update 11/12: postponed after CES. * AI (Tom): is there a prserv (ro) version? * Update 11/12: postponed after CES. * Jan-Simon to talk to Ryan hot to dump the sqlite from prserv * UPDATE 2019-01-22: Work resumed * master-next branch: * https://jira.automotivelinux.org/browse/SPEC-1837 * https://git.automotivelinux.org/AGL/meta-agl/commit/?h=next&id=cd14a879b91fb74e58c01e31248590ce5351c94e * ''repo grep AGL_FIXME'' * ''find . -name *AGLFIXME*'' * and you shall find the interesting spots to fix. * Update 13/11: no update * Update 27/11: no update * Update 11/12: no time, will look for a rebase and re-eval errors. * Update 2019-01-22: Scott / Tom will work on -next * Update 2019-02-05: * Scott: working on submodule fetcher upstream * Stephane: security manager fixed by Jose, dbus-cynara disabled (wgt install!) * smack ok, weston ok, no ivi-shell atm * binder + wgt builds work, issues around cynara * serves as reference for appfw update * issue with ptest packages, CI-build should enable agl-ptest * https://gerrit.automotivelinux.org/gerrit/17505 * tbd: selection of tests as some break the network filesystem * take board, run nbdroot and execute ptests one-by-one -> list of good/bad * Update 11/12: WIP * disabled systemd test as it disables network (bad in CI) * We'd have to disable any network test in the testsuite * Update 11/12: no update * AI for baylibre to fix systemd test * SPEC-1902, Stephane will do the reference features, Jan-Simon the CI features. * Update 27/11: WIP (draft in gerrit) * Update 11/12: WIP (draft in gerrit) * Update 2019-01-22: still draft in gerrit. * Update 2019-02-05: no time * Kusakabe-san: manual testing done by Fujitsu - which usb hardware for radio/BT if not built-in ? * https://wiki.automotivelinux.org/agl-distro/ces-2018-demo * radio/sdr: https://www.rtl-sdr.com/buy-rtl-sdr-dvb-t-dongles/ * bluetooth: * Product: CSR8510 A10 * AI: Walt: sample hardware list * Frederic: * lava-docker version of master: 19fe02f7a341e302e3ad560fcb4cc872f2201a5c * working on 64bit rpi3 filesystem New: * Jira-Zephyr vs Kiwi vs ??? * Kiwi: * + attach artifacts to the test (stored in kiwi) * repo manifest -r * + formatted output -> pdf * Zephyr: * create dummy set , plan and ad-hoc and planned execution * duplicate build+test+core UI plans from Konsulko to compare (Tom) * Talk about layer structure and moving components into the core * proposal update to thud first, then take a look at all components (add/move/remove) * meta-agl-devel/* * meta-oic ==== Meeting 2019/01/22 ==== ===Attendees=== Jan-Simon, Frederic, Kusakabe-san ===Topics=== * Test case submissions to JIRA * Test cases can be defined in Jira now. * Example: https://jira.automotivelinux.org/browse/SPEC-2132 * Todo: document tests executed * AI: Fujitsu & Konsulko - document test cases in JIRA * Fuego / Lava support * Patch for LAVA pending review https://git.lavasoftware.org/lava/lava/merge_requests/330 * * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * PRSERV setup in infra, addition to jobs pending * Update 13/11: prserv setup, need to use in jobs * Update 11/12: Jobs rework to use PRSERV (but will invalidate sstate-cache) * Update 11/12: postponed after CES. * AI (jsm): bitbake package-index * Update 11/12: postponed after CES. * AI (jsm): sync-out the rpm feeds for the snapshot builds * Update 11/12: more space on download server requested * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * Update 11/12: postponed after CES. * AI (Tom): is there a prserv (ro) version? * Update 11/12: postponed after CES. * UPDATE 2019-01-22: Work resumed * master-next branch: * https://jira.automotivelinux.org/browse/SPEC-1837 * https://git.automotivelinux.org/AGL/meta-agl/commit/?h=next&id=cd14a879b91fb74e58c01e31248590ce5351c94e * ''repo grep AGL_FIXME'' * ''find . -name *AGLFIXME*'' * and you shall find the interesting spots to fix. * Update 13/11: no update * Update 27/11: no update * Update 11/12: no time, will look for a rebase and re-eval errors. * Update 2019-01-22: Scott / Tom will work on -next * issue with ptest packages, CI-build should enable agl-ptest * https://gerrit.automotivelinux.org/gerrit/17505 * tbd: selection of tests as some break the network filesystem * take board, run nbdroot and execute ptests one-by-one -> list of good/bad * Update 27/11: * reviewed smack test and disabled it (meaningless) * LTP has meaningful smack tests * https://github.com/linux-test-project/ltp/tree/master/testcases/kernel/security/smack * Stephane: need a check for our default config being active * Who writes it ? (lava qa-testdefinition yaml needed) *Jose/Stephane* , Jan-Simon does the yaml * Update 11/12: WIP * disabled systemd test as it disables network (bad in CI) * We'd have to disable any network test in the testsuite * Update 11/12: no update * AI for baylibre to fix systemd test * SPEC-1902, Stephane will do the reference features, Jan-Simon the CI features. * Update 27/11: WIP (draft in gerrit) * Update 11/12: WIP (draft in gerrit) * Update 2019-01-22: still draft in gerrit. New: * Kusakabe-san: manual testing done by Fujitsu - which usb hardware for radio/BT if not built-in ? * https://wiki.automotivelinux.org/agl-distro/ces-2018-demo * radio/sdr: https://www.rtl-sdr.com/buy-rtl-sdr-dvb-t-dongles/ * bluetooth: * Product: CSR8510 A10 * AI: Walt: sample hardware list * Frederic: * lava-docker version of master: 19fe02f7a341e302e3ad560fcb4cc872f2201a5c * ==== Meeting 2018/12/11 ==== ===Attendees=== Jan-Simon, Frederic, Stephane, Li, Ronan * Fujitsu will do bi-weekly tests: * application tests: * bluetooth, hfp * radio * mediaplayer * navigation * hvac * dashboard * report by ? -> Jan-Simon to talk with Walt. (currently email) * Reviewing test report xls * Question about test result submission * We have now the 'zephyr' plugin for jira for test management installed * {{:eg-ciat:agl_app_test_case_20181127.xls| Test Report from Fujitsu}} * Update 11/12: * This week GG RC2 6.99.2 (report to be added) * Issues on: * Bluetooth (add report to jira -> Matt R) * Navigation (subsequent starts still bring not navigation to foreground) (->Tanikawa-san) * Mediaplayer (if usb-drive is removed, then mediaplayer hangs/does not play anymore, loop unstable) (-> Matt R) * {{:eg-ciat:agl_app_test_case_20181207.xls|Test Report from Fujitsu for 6.99.2 / GG RC2}} * Need to check if jira's are filed for these. * Fuego * Update 02/10 No update * Update 30/10: Enhanced Lava support patch posted https://lists.linuxfoundation.org/pipermail/fuego/2018-September/002422.html * Update 13/11: resend patch ? Li will send reminder * Update 27/11: Tim Bird looking at patch * Update 11/12: Tim is reviewing patch queue for a new release * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * PRSERV setup in infra, addition to jobs pending * Update 13/11: prserv setup, need to use in jobs * Update 11/12: Jobs rework to use PRSERV (but will invalidate sstate-cache) * Update 11/12: postponed after CES. * AI (jsm): bitbake package-index * Update 11/12: postponed after CES. * AI (jsm): sync-out the rpm feeds for the snapshot builds * Update 11/12: more space on download server requested * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * Update 11/12: postponed after CES. * AI (Tom): is there a prserv (ro) version? * Update 11/12: postponed after CES. * master-next branch: * https://jira.automotivelinux.org/browse/SPEC-1837 * https://git.automotivelinux.org/AGL/meta-agl/commit/?h=next&id=cd14a879b91fb74e58c01e31248590ce5351c94e * ''repo grep AGL_FIXME'' * ''find . -name *AGLFIXME*'' * and you shall find the interesting spots to fix. * Update 13/11: no update * Update 27/11: no update * Update 11/12: no time, will look for a rebase and re-eval errors. * issue with ptest packages, CI-build should enable agl-ptest * https://gerrit.automotivelinux.org/gerrit/17505 * tbd: selection of tests as some break the network filesystem * take board, run nbdroot and execute ptests one-by-one -> list of good/bad * Update 27/11: * reviewed smack test and disabled it (meaningless) * LTP has meaningful smack tests * https://github.com/linux-test-project/ltp/tree/master/testcases/kernel/security/smack * Stephane: need a check for our default config being active * Who writes it ? (lava qa-testdefinition yaml needed) *Jose/Stephane* , Jan-Simon does the yaml * Update 11/12: WIP * disabled systemd test as it disables network (bad in CI) * We'd have to disable any network test in the testsuite * Update 11/12: no update * SPEC-1902, Stephane will do the reference features, Jan-Simon the CI features. * Update 27/11: WIP (draft in gerrit) * Update 11/12: WIP (draft in gerrit) * INFRA updates * AGL core infra * LAVA * Lab iot/Lorient (France) - https://github.com/kernelci/lava-docker/blob/master/README.md * need public key of lab and deploy in lava master docker config * name of worker node (hostname -f): AGL-Lab-iotbzh * run *lavacli* to configure your worker + device + device_type Next call: Jan 15th 2019 , Jan 22nd 2019 ==== Meeting 2018/11/13 ==== ===Attendees=== Walt, Jan-Simon, Xiaoming Li, Ronan, Stephane, Frederic * Fujitsu will do bi-weekly tests: * application tests: * bluetooth, hfp * radio * mediaplayer * navigation * hvac * dashboard * report by ? -> Jan-Simon to talk with Walt. (currently email) * Reviewing test report xls * Question about test result submission * We have now the 'zephyr' plugin for jira for test management installed * {{:eg-ciat:agl_app_test_case_20181127.xls| Test Report from Fujitsu}} * Fuego * Update 02/10 No update * Update 30/10: Enhanced Lava support patch posted https://lists.linuxfoundation.org/pipermail/fuego/2018-September/002422.html * Update 13/11: resend patch ? Li will send reminder * Update 27/11: Tim Bird looking at patch * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * PRSERV setup in infra, addition to jobs pending * Update 13/11: prserv setup, need to use in jobs * AI (jsm): bitbake package-index * Update 30/10: no update * Update 13/11: no update * AI (jsm): sync-out the rpm feeds for the snapshot builds * Update 30/10: no update * Update 13/11: no update * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * Update 30/10: no update, waiting for feed * AI (Tom): is there a prserv (ro) version? * Update 30/10: no update * Update 13/11: no update * Update 27/11: no update * master-next branch: * https://jira.automotivelinux.org/browse/SPEC-1837 * https://git.automotivelinux.org/AGL/meta-agl/commit/?h=next&id=cd14a879b91fb74e58c01e31248590ce5351c94e * ''repo grep AGL_FIXME'' * ''find . -name *AGLFIXME*'' * and you shall find the interesting spots to fix. * Update 13/11: no update * Update 27/11: no update * issue with ptest packages, CI-build should enable agl-ptest * https://gerrit.automotivelinux.org/gerrit/17505 * tbd: selection of tests as some break the network filesystem * take board, run nbdroot and execute ptests one-by-one -> list of good/bad * Update 27/11: * reviewed smack test and disabled it (meaningless) * LTP has meaningful smack tests * https://github.com/linux-test-project/ltp/tree/master/testcases/kernel/security/smack * Stephane: need a check for our default config being active * Who writes it ? (lava qa-testdefinition yaml needed) *Jose/Stephane* , Jan-Simon does the yaml * disabled systemd test as it disables network (bad in CI) * We'd have to disable any network test in the testsuite * SPEC-1902, Stephane will do the reference features, Jan-Simon the CI features. * Update 27/11: WIP (draft in gerrit) ==== Meeting 2018/11/13 ==== ===Attendees=== Jan-Simon, Xiaoming Li, Ronan, Stephane, Frederic, Kusakabe-san Topics: * INFRA updates * AGL core infra * LAVA * AGL-core-lab (Germany) * Lab iot/Lorient (France) - https://github.com/kernelci/lava-docker/blob/master/README.md * need public key of lab and deploy in lava master docker config * name of worker node (hostname -f): AGL-Lab-iotbzh * Lab baylibre (US) * KernelCI * Performance issues * Navigation not user-friendly * 13/11: no update * Fuego * Update 02/10 No update * Update 30/10: Enhanced Lava support patch posted https://lists.linuxfoundation.org/pipermail/fuego/2018-September/002422.html * Update 13/11: resend patch ? Li will send reminder * Fujitsu will do bi-weekly tests: * application tests: * bluetooth, hfp * radio * mediaplayer * navigation * hvac * dashboard * report by ? -> Jan-Simon to talk with Walt. (currently email) * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * PRSERV setup in infra, addition to jobs pending * Update 13/11: prserv setup, need to use in jobs * AI (jsm): bitbake package-index * Update 30/10: no update * Update 13/11: no update * AI (jsm): sync-out the rpm feeds for the snapshot builds * Update 30/10: no update * Update 13/11: no update * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * Update 30/10: no update, waiting for feed * AI (Tom): is there a prserv (ro) version? * Update 30/10: no update * Update 13/11: no update * Current issues: * [[https://jira.automotivelinux.org/issues/?jql=status%20in%20(Open%2C%20%22In%20Progress%22%2C%20Reopened)%20AND%20component%20in%20(%22Build%20and%20Tools%22%2C%20CIAT)|Filter]] * New nbd client (proposal by Dominig) - SPEC-1423 * https://gerrit.automotivelinux.org/gerrit/#/c/13445/ * As discussed during F2F in Karlsruhe: * nbd.v3=1 (Stephane) - add boot argument to switch between old and newer implementation * Update 2018-10-30: no update (need lava code change first) * Update 2018-11-13: on hold, as lava changes needed first * https://jira.automotivelinux.org/browse/SPEC-1667 * Update 2018-10-30: no news * Update 2018-11-13: no update * 2018-10-30: CIAT infra sprint with baylibre * https://jira.automotivelinux.org/browse/SPEC-1847 * New issues: * master-next branch: * https://jira.automotivelinux.org/browse/SPEC-1837 * https://git.automotivelinux.org/AGL/meta-agl/commit/?h=next&id=cd14a879b91fb74e58c01e31248590ce5351c94e * ''repo grep AGL_FIXME'' * ''find . -name *AGLFIXME*'' * and you shall find the interesting spots to fix. * Update 13/11: no update * issue with ptest packages, CI-build should enable agl-ptest * https://gerrit.automotivelinux.org/gerrit/17505 * tbd: selection of tests as some break the network filesystem * take board, run nbdroot and execute ptests one-by-one -> list of good/bad * Discussion about autobuild, cmake, $make, and location of resulting files * cleanup needed when moving to cmake module. * a) autobuild/(agl|linux)/autobuild (currently fallback qmake/make : make && make package) * b) location of wgt: build dir (${B}) , $(toplevel build dir)/*.wgt * 13/11: ci-management (app-ci) and aglwgt.bbclass need to temporarily support old/new world (with big bbwarn) * Then convert apps over and remove exception New issues: * Slaves fail to reach download server -> SDK -> app builds 'just' succeed * SPEC-1902, Stephane will do the reference features, Jan-Simon the CI features. Next meeting in 2 week ==== Meeting 2018/10/30 ==== ===Attendees=== Jan-Simon, Xiaoming Li, Ronan, Romain, Stephane Topics: * ATS (Automated Testing Summit) readout/report * https://elinux.org/Automated_Testing_Summit * INFRA updates * AGL core infra * LAVA * AGL-core-lab (Germany) * Lab iot/Lorient (France) - https://github.com/kernelci/lava-docker/blob/master/README.md * Lab baylibre (US) * KernelCI * Performance issues * Navigation not user-friendly * Fuego * Update 02/10 No update * Update 30/10: Enhanced Lava support patch posted https://lists.linuxfoundation.org/pipermail/fuego/2018-September/002422.html * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * PRSERV setup in infra, addition to jobs pending * AI (jsm): bitbake package-index * Update 30/10: no update * AI (jsm): sync-out the rpm feeds for the snapshot builds * Update 30/10: no update * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) https://download.automotivelinux.org/AGL/snapshot/master/latest// * Update 30/10: no update, waiting for feed * AI (Tom): is there a prserv (ro) version? * Update 30/10: no update * Current issues: * [[https://jira.automotivelinux.org/issues/?jql=status%20in%20(Open%2C%20%22In%20Progress%22%2C%20Reopened)%20AND%20component%20in%20(%22Build%20and%20Tools%22%2C%20CIAT)|Filter]] * New nbd client (proposal by Dominig) - SPEC-1423 * https://gerrit.automotivelinux.org/gerrit/#/c/13445/ * As discussed during F2F in Karlsruhe: * nbd.v3=1 (Stephane) - add boot argument to switch between old and newer implementation * Update 2018-10-30: no update (need lava code change first) * Ronan: how to avoid broken build env due to binary bsp update? * detect if binary driver artifacts mismatch * Update 2018-08-21: waiting for BSP * Update 2018-10-30: Done * SPEC-1378 container * Update 2018-07-10: no news * Update 2018-08-21: build still failing with connection issues * Update 2018-10-30: new cloud network fixed the issue for now. close. * https://jira.automotivelinux.org/browse/SPEC-1667 * Update 2018-10-30: no news * 2018-10-30: CIAT infra sprint with baylibre * https://jira.automotivelinux.org/browse/SPEC-1847 * New issues: * master-next branch: * https://jira.automotivelinux.org/browse/SPEC-1837 * https://git.automotivelinux.org/AGL/meta-agl/commit/?h=next&id=cd14a879b91fb74e58c01e31248590ce5351c94e * ''repo grep AGL_FIXME'' * ''find . -name *AGLFIXME*'' * and you shall find the interesting spots to fix. * issue with ptest packages, CI-build should enable agl-ptest * https://gerrit.automotivelinux.org/gerrit/17505 * Discussion about autobuild, cmake, $make, and location of resulting files * cleanup needed when moving to cmake module. * a) autobuild/(agl|linux)/autobuild (currently fallback qmake/make : make && make package) * b) location of wgt: build dir (${B}) , $(toplevel build dir)/*.wgt Next meeting in 2 week ==== Meeting 2018/10/02 ==== ===Attendees=== Jan-Simon, Stephane, Tom, Ronan Topics: * INFRA updates * AGL core infra * LAVA * AGL-core-lab (Germany) - uprev'ed with lava-docker * Lab iot/Lorient (France) - https://github.com/kernelci/lava-docker/blob/master/README.md * Lab baylibre (US) - uprev'ed with lava-docker * KernelCI * Performance issues * Navigation not user-friendly * Fuego * Update 02/10 No update * SPEC-920 * Fix in gerrit 16867 * further discussion * AIs: * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) * In the end have a multi-config build with "bitbake agl-world" * atm share sstate-cache but separate TMPDIR * https://www.yoctoproject.org/docs/2.4.3/mega-manual/mega-manual.html#platdev-building-targets-with-multiple-configurations * bitbake multiconfig:m3conf multiconfig:h3conf multiconfig:intelcorei7conf dummy-image * conf/m3conf.conf * conf/h3conf.conf * conf/intelcorei7conf.conf * start with core ... once done ... lock the SIGNATURES ? * Initial step, use agl-demo-platform-crosssdk as target (only active profile atm) * AI (Ryan): activate PRSERV for snapshot builds * AI (jsm): bitbake package-index * AI (jsm): sync-out the rpm feeds for the snapshot builds * AI (Ronan): can we prepopulate the rpm feed info in the target ? (add to sandbox builds ... e.g. user 'lastest' ) * AI (Ton): is there a prserv (ro) version? * Current issues: * [[https://jira.automotivelinux.org/issues/?jql=status%20in%20(Open%2C%20%22In%20Progress%22%2C%20Reopened)%20AND%20component%20in%20(%22Build%20and%20Tools%22%2C%20CIAT)|Filter]] * New nbd client (proposal by Dominig) - SPEC-1423 * https://gerrit.automotivelinux.org/gerrit/#/c/13445/ * As discussed during F2F in Karlsruhe: * nbd.v3=1 (Stephane) - add boot argument to switch between old and newer implementation * Update 2018-08-21: no update * Ronan: how to avoid broken build env due to binary bsp update? * detect if binary driver artifacts mismatch * Update 2018-08-21: waiting for BSP * SPEC-1378 container * Update 2018-07-10: no news * Update 2018-08-21: build still failing with connection issues * docker container jobs fail: * https://build.automotivelinux.org/view/jjb-weekly-jobs/job/snapshot-jjb-container-weekly/101/console * adding extra line of defense checking the file size SPEC-979 * Update 2018-06-26: WIP * Update 2018-07-10: debugged the container build, still broken * docker save docker.automotivelinux.org/agl/worker-generic:5.99-116 | xz -T0 -c >docker_agl_worker-generic-5.99-116.tar.xz * Error response from daemon: reference does not exist * generic container vs. xds container discussion: * generic container uncompressed: 1.5GB (~380MB compressed) * xds containner uncompressed: 600MB (~150MB compressed) * converge by cleaning up generic one ? * for the download server AI:jsmoeller query download stats * doc site has 2 different chapters: * http://docs.automotivelinux.org/docs/getting_started/en/dev/reference/setup-sdk-environment.html * http://docs.automotivelinux.org/docs/devguides/en/dev/reference/xds/part-1/2-1_install-xds-server-docker.html * different audiences, but bad guidance/nomenclature/naming aas to what should be used where * sleep over it and discuss further in SAT * Sebastian investigated the disk/space usage * 1) most taken by yocto dependencies (gcc) * 2) docs (nodejs) * 3) 100mb for x11 clients (rdp) * Use platform and app-development (native and html5) containers. * Smaller SDK size (e.g. no kernel-devsrc and so on) for app developers (jira for GG) * https://jira.automotivelinux.org/browse/SPEC-1667 * New issues: * ==== Meeting 2018/09/25 ==== ===Attendees=== Jan-Simon, Frederic, Kurokawa-san, Khiem Nguyen, Xiaoming Li, Ronan, Scott, Tom, 'Renesas dev', Topic: SPEC-920 ( https://jira.automotivelinux.org/browse/SPEC-920 ) Notes: * RPM feed for image and SDK by eg. dnf or dnf-sdk wrapper * Limitation/Problem observed: * PRSERV use the HASH to match-up with the revision (generated/stored/served) * Revert in metadata (recipe) will lead to 'old' revision being used (prserv replies with old rev) * https://github.com/iotbzh/YoMo * poky/meta/conf/bitbake.conf:PKGR ?= "${PR}${EXTENDPRAUTO}" * https://github.com/iotbzh/YoMo/blob/master/doc/YoMo_repositories.md * Remaining problem: metadata revert will lead to r4->r3 due to hash being recognized by prserv as existing * AIs: * To try things out, get in 16667 to master only (it is optional and not enabled by default) * Scott: take a look at the prserv if we can solve above issue on rpm feeds * AI (jsm/ryan): forbid revert button (?edit button?) * Tom: may need policy around reverts and bumping PR value (at least for meta-layers) ==== Meeting 2018/08/21 ==== ===Attendees=== Jan-Simon, Ronan, Li Xiaoming, Frederic == AGL CIAT meeting == ===Topics=== * Status updates: * Infra: * slave lab in Lorient being setup https://lava.automotivelinux.org/scheduler/worker/AGL-Lab-iotbzh.lorient.iot * Further investigation needed based on Dominigs email. * Need to try pxe next * Update 2018-06-26: Input during ALS, working sample with ipxe shown. Doc to be sent to ML. * Liu showed procedure * Advanced Network stack enabled. * Enable network boot in bios boot config screen (Fixed boot order) * build IPXE from git.ipxe.org/ipxe.git ; cd ipxe/src * shell.ipxe: * #!ipxe * shell * make bin-x86_64-efi/ipxe.efi EMBED=./shell.ipxe * Frederic: got pxe shell and network boot * Update 2018-07-10: * ipxe works , but LAVA had issues with the serial (https://lava.automotivelinux.org/scheduler/job/5735#L179) * uprev for new lava server version pending when serial is debugged * Update 2018-07-24: * serial debugging still WIP * Update 2018-08-21: * need to follow-up * Lava * waiting for LAVA rev. 2018.04 to be ready in lava-docker * Branch available, see https://jira.automotivelinux.org/browse/SPEC-1299 * Means we also have to update the slaves (=new version of lava-dispatcher) * WIP: uprev slaves. * Update 2018-06-26: Schedule next week (week 27) for lab uprev. * Update 2018-07-10: Jan-Simon started to uprev the AGL-core-lab and uses the lava-docker * Issues: * certificate exchange was missing (added meanwhile) * WIP currently: Need to create new server certificate and exchange * http://lavadocker.automotivelinux.org:10080/ but needs fix for https * Update 2018-07-24: * fix zmq auth issue * persistent db * 2018.04 uprev done * Jan-Simon working on docker boards.yaml for master server * Update 2018-08-21: * Connection issues between slave and master, need to debug with Ryan. * KernelCI * webui changes WIP * tabs for gerrit/CI , weekly, daily, release builds * Update 2018-06-26: demo last week during ALS, plan to deploy week 27 * Update 2018-07-10: no update, delayed * Update 2018-07-24: baylibre has test instance deployed with ui changes * Update 2018-08-21: * Deploying latest changes * Fuego * any update ? - 2018-05-15: no updates * 2018-05-29: no update on lava support * Update 2018-06-26: Presentation during ALS with better integration (add link) * 2018-07-10: no patches for integration, yet. * 2018-07-24: no update * Update 2018-08-21: no update * Current issues: * [[https://jira.automotivelinux.org/issues/?jql=status%20in%20(Open%2C%20%22In%20Progress%22%2C%20Reopened)%20AND%20component%20in%20(%22Build%20and%20Tools%22%2C%20CIAT)|Filter]] * New nbd client (proposal by Dominig) - SPEC-1423 * https://gerrit.automotivelinux.org/gerrit/#/c/13445/ * As discussed during F2F in Karlsruhe: * nbd.v3=1 (Stephane) - add boot argument to switch between old and newer implementation * port new nbd-client for lava-dispatcher * Update 2018-05-15: need to investigate uprev of nbd-client in yocto/oe before we uprev. * Update 2018-07-24: investigating client versions * Update 2018-08-21: no update * Ronan: how to avoid broken build env due to binary bsp update? * detect if binary driver artifacts mismatch * separate dummy recipe/package with no rpm (empty) * which does the check for the binary at configure stage. * e.g. at configure step. * Update 2018-07-24: waiting for rc3 (for BSP 3.9) * Update 2018-08-21: waiting for BSP * SPEC-1378 container * Update 2018-07-10: no news * Update 2018-08-21: build still failing with connection issues * docker container jobs fail: * https://build.automotivelinux.org/view/jjb-weekly-jobs/job/snapshot-jjb-container-weekly/101/console * adding extra line of defense checking the file size SPEC-979 * Update 2018-06-26: WIP * Update 2018-07-10: debugged the container build, still broken * docker save docker.automotivelinux.org/agl/worker-generic:5.99-116 | xz -T0 -c >docker_agl_worker-generic-5.99-116.tar.xz * Error response from daemon: reference does not exist * generic container vs. xds container discussion: * generic container uncompressed: 1.5GB (~380MB compressed) * xds containner uncompressed: 600MB (~150MB compressed) * converge by cleaning up generic one ? * for the download server AI:jsmoeller query download stats * doc site has 2 different chapters: * http://docs.automotivelinux.org/docs/getting_started/en/dev/reference/setup-sdk-environment.html * http://docs.automotivelinux.org/docs/devguides/en/dev/reference/xds/part-1/2-1_install-xds-server-docker.html * different audiences, but bad guidance/nomenclature/naming aas to what should be used where * sleep over it and discuss further in SAT * Sebastian investigated the disk/space usage * 1) most taken by yocto dependencies (gcc) * 2) docs (nodejs) * 3) 100mb for x11 clients (rdp) * Use platform and app-development (native and html5) containers. * Smaller SDK size (e.g. no kernel-devsrc and so on) for app developers (jira for GG) * New issues: * https://jira.automotivelinux.org/browse/SPEC-1667 * ===Next meeting(s)=== Next call in 2 weeks, same place, same time (13:00 UTC) ==== Meeting 2018/07/24 ==== ===Attendees=== Jan-Simon, Frederic, Li Xiaoming, Clement, Kevin, Stephane, Ronan == AGL CIAT meeting == ===Topics=== * Status updates: * Infra: * slave lab in Lorient being setup https://lava.automotivelinux.org/scheduler/worker/AGL-Lab-iotbzh.lorient.iot * Further investigation needed based on Dominigs email. * Need to try pxe next * Update 2018-06-26: Input during ALS, working sample with ipxe shown. Doc to be sent to ML. * Liu showed procedure * Advanced Network stack enabled. * Enable network boot in bios boot config screen (Fixed boot order) * build IPXE from git.ipxe.org/ipxe.git ; cd ipxe/src * shell.ipxe: * #!ipxe * shell * make bin-x86_64-efi/ipxe.efi EMBED=./shell.ipxe * Frederic: got pxe shell and network boot * Update 2018-07-10: * ipxe works , but LAVA had issues with the serial (https://lava.automotivelinux.org/scheduler/job/5735#L179) * uprev for new lava server version pending when serial is debugged * Update 2018-07-24: * serial debugging still WIP * download server speed issues under investigation * for now we doubled the speed ... = ~1.5MB/s max * Debugging ongoing, workaround for CI WIP * Update 2018-06-26: Seems to be an upstream ISP issue. Workaround seems in place but no final Solution atm. * Update 2018-07-10: Cloud internal network structure changed (slaves on own private net now) * Update 2018-07-24: taking care of lab in Lorient. * Lava * waiting for LAVA rev. 2018.04 to be ready in lava-docker * Branch available, see https://jira.automotivelinux.org/browse/SPEC-1299 * Means we also have to update the slaves (=new version of lava-dispatcher) * WIP: uprev slaves. * Update 2018-06-26: Schedule next week (week 27) for lab uprev. * Update 2018-07-10: Jan-Simon started to uprev the AGL-core-lab and uses the lava-docker * Issues: * certificate exchange was missing (added meanwhile) * WIP currently: Need to create new server certificate and exchange * http://lavadocker.automotivelinux.org:10080/ but needs fix for https * Update 2018-07-24: * fix zmq auth issue * persistent db * 2018.04 uprev done * Jan-Simon working on docker boards.yaml for master server * KernelCI * webui changes WIP * tabs for gerrit/CI , weekly, daily, release builds * Update 2018-06-26: demo last week during ALS, plan to deploy week 27 * Update 2018-07-10: no update, delayed * Update 2018-07-24: baylibre has test instance deployed with ui changes * * Fuego * any update ? - 2018-05-15: no updates * 2018-05-29: no update on lava support * Update 2018-06-26: Presentation during ALS with better integration (add link) * 2018-07-10: no patches for integration, yet. * 2018-07-24: no update * Current issues: * [[https://jira.automotivelinux.org/issues/?jql=status%20in%20(Open%2C%20%22In%20Progress%22%2C%20Reopened)%20AND%20component%20in%20(%22Build%20and%20Tools%22%2C%20CIAT)|Filter]] * SPEC-1378 * Update 2018-05-29: wip * Update 2018-06-12: wip * Update 2018-06-26: Reduce build frequency to RC's final / ondemand / move proper container to release folder. * Update 2018-07-10: no news * New nbd client (proposal by Dominig) - SPEC-1423 * https://gerrit.automotivelinux.org/gerrit/#/c/13445/ * As discussed during F2F in Karlsruhe: * nbd.v3=1 (Stephane) - add boot argument to switch between old and newer implementation * port new nbd-client for lava-dispatcher * Update 2018-05-15: need to investigate uprev of nbd-client in yocto/oe before we uprev. * Update 2018-05-29: no update * Update 2018-06-12: investigating client versions * Update 2018-06-26: investigating client versions * Update 2018-07-10: investigating client versions * Update 2018-07-24: investigating client versions * Ronan: how to avoid broken build env due to binary bsp update? * detect if binary driver artifacts mismatch * separate dummy recipe/package with no rpm (empty) * which does the check for the binary at configure stage. * e.g. at configure step. * Update 2018-06-12: no update * Update 2018-06-26: no update, plan is to do a python class. * Update 2018-07-10: investigating for new bsp (3.9) * Update 2018-07-24: waiting for rc3 (for BSP 3.9) * docker container jobs fail: * https://build.automotivelinux.org/view/jjb-weekly-jobs/job/snapshot-jjb-container-weekly/101/console * adding extra line of defense checking the file size SPEC-979 * Update 2018-06-26: WIP * Update 2018-07-10: debugged the container build, still broken * docker save docker.automotivelinux.org/agl/worker-generic:5.99-116 | xz -T0 -c >docker_agl_worker-generic-5.99-116.tar.xz * Error response from daemon: reference does not exist * generic container vs. xds container discussion: * generic container uncompressed: 1.5GB (~380MB compressed) * xds containner uncompressed: 600MB (~150MB compressed) * converge by cleaning up generic one ? * for the download server AI:jsmoeller query download stats * doc site has 2 different chapters: * http://docs.automotivelinux.org/docs/getting_started/en/dev/reference/setup-sdk-environment.html * http://docs.automotivelinux.org/docs/devguides/en/dev/reference/xds/part-1/2-1_install-xds-server-docker.html * different audiences, but bad guidance/nomenclature/naming aas to what should be used where * sleep over it and discuss further in SAT * Sebastian investigated the disk/space usage * 1) most taken by yocto dependencies (gcc) * 2) docs (nodejs) * 3) 100mb for x11 clients (rdp) * Use platform and app-development (native and html5) containers. * New issues: * Smaller SDK size (e.g. no kernel-devsrc and so on) for app developers (jira for GG) ===Next meeting(s)=== Next call in 2 weeks, same place, same time (13:00 UTC) ==== Meeting 2018/07/10 ==== ===Attendees=== Jan-Simon, Frederic, Ronan, Tanibata-san, Li Xiaoming == AGL CIAT meeting == ===Topics=== * Status updates: * Infra: * slave lab in Lorient being setup https://lava.automotivelinux.org/scheduler/worker/AGL-Lab-iotbzh.lorient.iot * Further investigation needed based on Dominigs email. * Need to try pxe next * Update 2018-06-26: Input during ALS, working sample with ipxe shown. Doc to be sent to ML. * Liu showed procedure * Advanced Network stack enabled. * Enable network boot in bios boot config screen (Fixed boot order) * build IPXE from git.ipxe.org/ipxe.git ; cd ipxe/src * shell.ipxe: * #!ipxe * shell * make bin-x86_64-efi/ipxe.efi EMBED=./shell.ipxe * Frederic: got pxe shell and network boot * Update 2018-07-10: * ipxe works , but LAVA had issues with the serial (https://lava.automotivelinux.org/scheduler/job/5735#L179) * uprev for new lava server version pending when serial is debugged. * download server speed issues under investigation * for now we doubled the speed ... = ~1.5MB/s max * Debugging ongoing, workaround for CI WIP * Update 2018-06-26: Seems to be an upstream ISP issue. Workaround seems in place but no final Solution atm. * Update 2018-07-10: Cloud internal network structure changed (slaves on own private net now) * Lava * waiting for LAVA rev. 2018.04 to be ready in lava-docker * Branch available, see https://jira.automotivelinux.org/browse/SPEC-1299 * Means we also have to update the slaves (=new version of lava-dispatcher) * WIP: uprev slaves. * Update 2018-06-26: Schedule next week (week 27) for lab uprev. * Update 2018-07-10: Jan-Simon started to uprev the AGL-core-lab and uses the lava-docker * Issues: * certificate exchange was missing (added meanwhile) * WIP currently: Need to create new server certificate and exchange * http://lavadocker.automotivelinux.org:10080/ but needs fix for https * KernelCI * webui changes WIP * tabs for gerrit/CI , weekly, daily, release builds * Update 2018-06-26: demo last week during ALS, plan to deploy week 27 * Update 2018-07-18: no update, delayed * Fuego * any update ? - 2018-05-15: no updates * 2018-05-29: no update on lava support * Update 2018-06-26: Presentation during ALS with better integration (add link) * 2018-07-10: no patches for integration, yet. * Current issues: * [[https://jira.automotivelinux.org/issues/?jql=status%20in%20(Open%2C%20%22In%20Progress%22%2C%20Reopened)%20AND%20component%20in%20(%22Build%20and%20Tools%22%2C%20CIAT)|Filter]] * SPEC-1378 * Update 2018-05-29: wip * Update 2018-06-12: wip * Update 2018-06-26: Reduce build frequency to RC's final / ondemand / move proper container to release folder. * Update 2018-07-10: no news * New nbd client (proposal by Dominig) - SPEC-1423 * https://gerrit.automotivelinux.org/gerrit/#/c/13445/ * As discussed during F2F in Karlsruhe: * nbd.v3=1 (Stephane) - add boot argument to switch between old and newer implementation * port new nbd-client for lava-dispatcher * Update 2018-05-15: need to investigate uprev of nbd-client in yocto/oe before we uprev. * Update 2018-05-29: no update * Update 2018-06-12: investigating client versions * Update 2018-06-26: investigating client versions * Update 2018-07-10: investigating client versions * Ronan: how to avoid broken build env due to binary bsp update? * detect if binary driver artifacts mismatch * separate dummy recipe/package with no rpm (empty) * which does the check for the binary at configure stage. * e.g. at configure step. * Update 2018-06-12: no update * Update 2018-06-26: no update, plan is to do a python class. * Update 2018-07-10: investigating for new bsp (3.9) * docker container jobs fail: * https://build.automotivelinux.org/view/jjb-weekly-jobs/job/snapshot-jjb-container-weekly/101/console * adding extra line of defense checking the file size SPEC-979 * Update 2018-06-26: WIP * Update 2018-07-10: debugged the container build, still broken * docker save docker.automotivelinux.org/agl/worker-generic:5.99-116 | xz -T0 -c >docker_agl_worker-generic-5.99-116.tar.xz * Error response from daemon: reference does not exist * New issues: * widget creation workflow (SPEC-1579) - workflow was done in '16 and documented here: * https://wiki.automotivelinux.org/troubleshooting/app-recipes?s[]=make%20package * need to avoid breaking old releases going forward * app builds with SDK in CI and aglwgt.bbclass * https://jira.automotivelinux.org/browse/SPEC-1530 * Ronan pushes a fix to gerrit (cmake module) * Upstream ?! AI: Ronan to open jira to track and do the submit upstream * will use patch as short-term fix until submitted upstream * Go stdlib install. * generic container vs. xds container discussion: * generic container uncompressed: 1.5GB (~380MB compressed) * xds containner uncompressed: 600MB (~150MB compressed) * converge by cleaning up generic one ? * for the download server AI:jsmoeller query download stats * doc site has 2 different chapters: * http://docs.automotivelinux.org/docs/getting_started/en/dev/reference/setup-sdk-environment.html * http://docs.automotivelinux.org/docs/devguides/en/dev/reference/xds/part-1/2-1_install-xds-server-docker.html * different audiences, but bad guidance/nomenclature/naming aas to what should be used where * sleep over it and discuss further in SAT ===Next meeting(s)=== Next call in 2 weeks, same place, same time (13:00 UTC) ==== Meeting 2018/06/26 ==== ===Attendees=== Jan-Simon, Frederic, Ronan, Stephane, Liu, Oki-san == AGL CIAT meeting == ===Topics=== * Status updates: * Infra: * slave lab in Lorient being setup https://lava.automotivelinux.org/scheduler/worker/AGL-Lab-iotbzh.lorient.iot * Further investigation needed based on Dominigs email. * Need to try pxe next * Update 2018-06-26: Input during ALS, working sample with ipxe shown. Doc to be sent to ML. * Liu showed procedure * Advanced Network stack enabled. * Enable network boot in bios boot config screen (Fixed boot order) * build IPXE from git.ipxe.org/ipxe.git ; cd ipxe/src * shell.ipxe: * #!ipxe * shell * make bin-x86_64-efi/ipxe.efi EMBED=./shell.ipxe * Frederic: got pxe shell and network boot * download server speed issues under investigation * for now we doubled the speed ... = ~1.5MB/s max * Debugging ongoing, workaround for CI WIP * Update 2018-06-26: Seems to be an upstream ISP issue. Workaround seems in place but no final Solution atm. * Lava * waiting for LAVA rev. 2018.04 to be ready in lava-docker * Branch available, see https://jira.automotivelinux.org/browse/SPEC-1299 * Means we also have to update the slaves (=new version of lava-dispatcher) * WIP: uprev slaves. * Update 2018-06-26: Schedule next week (week 27) for lab uprev. * KernelCI * webui changes WIP * tabs for gerrit/CI , weekly, daily, release builds * * Update 2018-06-26: demo last week during ALS, plan to deploy week 27 * Fuego * any update ? - 2018-05-15: no updates * 2018-05-29: no update on lava support * Update 2018-06-26: Presentation during ALS with better integration (add link) * Current issues: * [[https://jira.automotivelinux.org/issues/?jql=status%20in%20(Open%2C%20%22In%20Progress%22%2C%20Reopened)%20AND%20component%20in%20(%22Build%20and%20Tools%22%2C%20CIAT)|Filter]] * SPEC-1378 * Update 2018-05-29: wip * Update 2018-06-12: wip * Update 2018-06-26: Reduce build frequency to RC's final / ondemand / move proper container to release folder. * New nbd client (proposal by Dominig) - SPEC-1423 * https://gerrit.automotivelinux.org/gerrit/#/c/13445/ * As discussed during F2F in Karlsruhe: * nbd.v3=1 (Stephane) - add boot argument to switch between old and newer implementation * port new nbd-client for lava-dispatcher * Update 2018-05-15: need to investigate uprev of nbd-client in yocto/oe before we uprev. * Update 2018-05-29: no update * Update 2018-06-12: investigating client versions * Update 2018-06-26: investigating client versions * Ronan: how to avoid broken build env due to binary bsp update? * detect if binary driver artifacts mismatch * separate dummy recipe/package with no rpm (empty) * which does the check for the binary at configure stage. * e.g. at configure step. * Update 2018-06-12: no update * Update 2018-06-26: no update, plan is to do a python class. * https://jira.automotivelinux.org/browse/SPEC-1434 * Update 2018-06-26: Closed. Fix in current master. * docker container jobs fail: * https://build.automotivelinux.org/view/jjb-weekly-jobs/job/snapshot-jjb-container-weekly/101/console * adding extra line of defense checking the file size SPEC-979 * Update 2018-06-26: WIP * New issues: * no ===Next meeting(s)=== Next call in 2 weeks, same place, same time (13:00 UTC) ==== Meeting 2018/05/29 ==== ===Attendees=== Jan-Simon, == AGL CIAT meeting == ===Topics=== * Status updates: * Infra: * slave lab in Lorient being setup https://lava.automotivelinux.org/scheduler/worker/AGL-Lab-iotbzh.lorient.iot * worked during f2f on netboot with uefi, need to either use grubefi or pxe . * Further investigation needed based on Dominigs email. * Need to try pxe next * download server speed issues under investigation * for now we doubled the speed ... = ~1.5MB/s max * Ryan to does a debug session with provider * Debugging ongoing, workaround for CI WIP * Lava * waiting for LAVA rev. 2018.04 to be ready in lava-docker * Branch available, see https://jira.automotivelinux.org/browse/SPEC-1299 * Means we also have to update the slaves (=new version of lava-dispatcher) * WIP: uprev slaves. * KernelCI * webui changes WIP * tabs for gerrit/CI , weekly, daily, release builds * Fuego * any update ? - 2018-05-15: no updates * 2018-05-29: no update on lava support * Current issues: * [[https://jira.automotivelinux.org/issues/?jql=status%20in%20(Open%2C%20%22In%20Progress%22%2C%20Reopened)%20AND%20component%20in%20(%22Build%20and%20Tools%22%2C%20CIAT)|Filter]] * SPEC-1378 * Update 2018-05-29: wip * Update 2018-06-12: wip * New nbd client (proposal by Dominig) - SPEC-1423 * https://gerrit.automotivelinux.org/gerrit/#/c/13445/ * As discussed during F2F in Karlsruhe: * nbd.v3=1 (Stephane) - add boot argument to switch between old and newer implementation * port new nbd-client for lava-dispatcher * Update 2018-05-15: need to investigate uprev of nbd-client in yocto/oe before we uprev. * Update 2018-05-29: no update * Update 2018-06-12: investigating client versions * Ronan: how to avoid broken build env due to binary bsp update? * detect if binary driver artifacts mismatch * separate dummy recipe/package with no rpm (empty) * which does the check for the binary at configure stage. * e.g. at configure step. * Update 2018-06-12: no update * https://jira.automotivelinux.org/browse/SPEC-1434 * New issues: * docker container jobs fail: * https://build.automotivelinux.org/view/jjb-weekly-jobs/job/snapshot-jjb-container-weekly/101/console * adding extra line of defense checking the file size SPEC-979 ===Next meeting(s)=== Next call in 2 weeks, same place, same time (13:00 UTC) ==== Meeting 2018/05/29 ==== ===Attendees=== Jan-Simon, Frederic Marec, == AGL CIAT meeting == ===Topics=== * Status updates: * Infra: * slave lab in Lorient being setup https://lava.automotivelinux.org/scheduler/worker/AGL-Lab-iotbzh.lorient.iot * healthcheck jobs need debugging * https://lava.automotivelinux.org/scheduler/job/5556 * scheduled for wed. * download server speed issues under investigation * for now we doubled the speed ... = ~1.5MB/s max * Ryan to does a debug session with provider * Lava * waiting for LAVA rev. 2018.04 to be ready in lava-docker * Branch available, see https://jira.automotivelinux.org/browse/SPEC-1299 * Means we also have to update the slaves (=new version of lava-dispatcher) * KernelCI * webui changes WIP * tabs for gerrit/CI , weekly, daily, release builds * Fuego * any update ? - 2018-05-15: no updates * 2018-05-29: no update on lava support * Current issues: * [[https://jira.automotivelinux.org/issues/?jql=status%20in%20(Open%2C%20%22In%20Progress%22%2C%20Reopened)%20AND%20component%20in%20(%22Build%20and%20Tools%22%2C%20CIAT)|Filter]] * SPEC-1378 * Update 2018-04-17: No update * EE & FF * Update 2018-05-15: wip, creating templates fir ci-management repo * Update 2018-05-29: wip * New nbd client (proposal by Dominig) - SPEC-1423 * https://gerrit.automotivelinux.org/gerrit/#/c/13445/ * As discussed during F2F in Karlsruhe: * nbd.v3=1 (Stephane) - add boot argument to switch between old and newer implementation * port new nbd-client for lava-dispatcher * Update 2018-05-15: need to investigate uprev of nbd-client in yocto/oe before we uprev. * Update 2018-05-29: no update * Ronan: how to avoid broken build env due to binary bsp update? * detect if binary driver artifacts mismatch * separate dummy recipe/package with no rpm (empty) * which does the check for the binary at configure stage. * e.g. at configure step. * New issues: * Layer reorg in stage 3. * https://jira.automotivelinux.org/browse/SPEC-1434 ===Next meeting(s)=== Next call in 2 weeks, same place, same time (13:00 UTC) ==== Meeting 2018/05/15 ==== ===Attendees=== Jan-Simon, Frederic Marec, Liu Wenlong, Tanibata-san, Oki-san, Ronan Le Martret, Stephane Desneux, Tiejun Chen ===Topics=== == AGL CIAT meeting == * Status updates: * Infra: * slave lab in Lorient being setup https://lava.automotivelinux.org/scheduler/worker/AGL-Lab-iotbzh.lorient.iot * download server speed issues under investigation * It thinks it is an issue within the openstack network * Lava * waiting for LAVA rev. 2018.4 to be ready in lava-docker * KernelCI * webui changes WIP * tabs for gerrit/CI , weekly, daily, release builds * Fuego * any update ? - 2018-05-15: no updates * Current issues: * [[https://jira.automotivelinux.org/issues/?jql=status%20in%20(Open%2C%20%22In%20Progress%22%2C%20Reopened)%20AND%20component%20in%20(%22Build%20and%20Tools%22%2C%20CIAT)|Filter]] * SPEC-979 (Stephane): * docker images: See jira. Fixing docker build and using BUILD_NUMBER. Also separate release job for container images (other from weekly container build). * SPEC-1378 * Update 2018-04-17: No update * EE & FF * Update 2018-05-15: wip, creating templates fir ci-management repo * New nbd client (proposal by Dominig) - SPEC-1423 * https://gerrit.automotivelinux.org/gerrit/#/c/13445/ * As discussed during F2F in Karlsruhe: * nbd.v3=1 (Stephane) - add boot argument to switch between old and newer implementation * port new nbd-client for lava-dispatcher * Update 2018-05-15: need to investigate uprev of nbd-client in yocto/oe before we uprev. * m3ulcb, h3ulcb build issues in jira, Ronan is looking into it. * Update 2018-05-15: done * Ronan: how to avoid broken build env due to binary bsp update? * detect if binary driver artifacts mismatch * New issues: * Layer reorg in stage 3. ===Next meeting(s)=== Next call in 2 weeks, same place, same time (13:00 UTC) ==== Meeting 2018/04/17 ==== == AGL CIAT meeting == * Readout of the [[https://wiki.automotivelinux.org/agl-distro/apr2018-f2f|Karlsruhe F2F Meeting]]: * General CIAT Session ([[https://docs.google.com/presentation/d/1u1ezfkA9FUisPMTAZDcsuBmPsSLWtvJXpWnr3PiRFUk/edit?usp=sharing|Slides]]) * Fuego subsession ([[https://docs.google.com/presentation/d/1faeQzB4YolsodrHX52wiyAYMvwvLBxheNg9zgx6pnC0/edit?usp=sharing|Slides1]], [[https://drive.google.com/file/d/0BxAiYws0XCAxWDMxanRRdlV3cUNuMldSWEZ2S3FYQkotVmd3/view?usp=sharing|Slides2]]) * Status updates: * Infra: blocked due to travel/F2F, continuing this week * Current issues: * SPEC-1374: tried to reduce build parallelism * Update 2018-04-17: Fixed. Was problem in cogent repo - Ronan fixed by blacklisting. * SPEC-1334: https://jira.automotivelinux.org/browse/SPEC-1334 * infra to run ptests is there * Update 2018-04-17: Fixed by Jose. ptests can be built again. * SPEC-781: * Missing license file checksum * Update 2018-04-17: Fixed. Closed. * https://git.automotivelinux.org/ci-management/tree/packer <- jenkins build slaves * install 'packer' * then run ./local_packer_run.sh * image is in 'OUTPUT' can be run in qemu * SPEC-979 (Stephane): * docker images: See jira. Fixing docker build and using BUILD_NUMBER. Also separate release job for container images (other from weekly container build). * Created * SPEC-1378 * Update 2018-04-17: No update * SPEC-1379 * Update 2018-04-17: Merged. Closed. * New issues: * New nbd client (proposal by Dominig) * https://gerrit.automotivelinux.org/gerrit/#/c/13445/ * As discussed during F2F in Karlsruhe: * nbd.v3=1 (Stephane) - add boot argument to switch between old and newer implementation * port new nbd-client for lava-dispatcher * m3ulcb, h3ulcb build issues in jira, Ronan is looking into it. * Ronan: how to avoid broken build env due to binary bsp update? ===Next meeting(s)=== Next call in 4 weeks (due to may 1st being a public holiday) same place, same time (13:00 UTC) ==== Meeting 2018/04/03 ==== ===Attendees=== Jan-Simon, Kurokawa-san, Kusakabe-san, Stephane, Oki-san, Ronan, Frederic, Liu Wenlong, Mohammed Gueroud, Vinod, ===Topics=== == AGL CIAT meeting == * Status updates: * KernelCI * configuration in progress * LAVA docker in LFIT * setting ACL and tokens * CIBT / LAVA * https://jira.automotivelinux.org/browse/SPEC-1209 * Update: font package is causing the issue (Kudos to Jose) * Plan: make this update non-blocking for serial terminal access * WIP (enable KVM + fix to postinst). * Update 2018-04-03: KVM active in lava lab, thus mitigated. * Next actions: enabling more ptests and test scripts * New: * SPEC-1374: tried to reduce build parallelism * SPEC-1334: https://jira.automotivelinux.org/browse/SPEC-1334 * infra to run ptests is there, * SPEC-781: * Missing license file checksum * https://git.automotivelinux.org/ci-management/tree/packer <- jenkins build slaves * install 'packer' * then run ./local_packer_run.sh * image is in 'OUTPUT' can be run in qemu * SPEC-979 (Stephane): * docker images: See jira. Fixing docker build and using BUILD_NUMBER. Also separate release job for container images (other from weekly container build). * Created * SPEC-1378 * SPEC-1379 * SPEC-1380 * New board support in FUEGO (Oki-san) * Introducing during F2F in Karlsuhe on Friday ===Next meeting(s)=== Next call in 2 weeks (apr 17th) same place, same time (13:00 UTC) ==== Meeting 2018/03/06 ==== ===Attendees=== Jan-Simon Möller, Frederic Marec ===Topics=== * Status updates: * Infra: * downtime required today to re-setup firewall node in cloud env (to improve BW) * See email. * KernelCI * deployed * remaining issue: access to data older than 14d (UI) * LAVA docker in LFIT * restoring database dump * setting ACL and tokens * CIBT / LAVA * https://jira.automotivelinux.org/browse/SPEC-1209 * Update: font package is causing the issue * Plan: make this update non-blocking for serial terminal access * WIP (enable KVM + fix to postinst). * Lab-in-a-box * Training session during AMM - update/feedback ? ===Next meeting(s)=== Next call in 2 weeks (march 20th) same place, same time (13:00 UTC) ==== Meeting 2018/02/13 ==== ===Attendees=== Jan-Simon, Tim, Frederic, Kyohhei-san, Ronan, Kusakabe-san, Stephane ===Topics=== == Hosting the FUEGO meeting today, welcome! == * Status update * Fuego status * next Major release areas of focus * Documentation conversion to reStructuredText * Documentation for end users (not developers) * How to customize tests (criteria.json) * Dynamic test documentation in reStructuredText * Combination of human-readable and Fuego-generated content * Fuego automated release testing (Fuego selftest) * ProFusion has started work * Have done major cleanups to Dockerfile * Working towards having a downloadable docker image (not requiring image build by end user) * work is in fuego 'next' branch of tbird20d git repository * System provisioning * Build and deploy system to target * Stuck working on some Board Farm management software issues * May add following soon: * board status check * board hardware reboot * board serial monitor (maybe) * Work is currently being done in the 'master' branch * A few tests changes recently: * added more Realtime tests: backfire, migratetest, deadlinetest * more changes for LTP: parsing of realtime results == AGL CIAT meeting == * Status updates: * KernelCI * https termination issue resolved, deploying now. * https://github.com/lucj/kernelci-docker/issues/13#issuecomment-360921103 * LAVA docker in LFIT * working on deployment with external labs * migrating the existing configuration * CIBT / LAVA * need helping hand debugging qemux86-64 target * https://build.automotivelinux.org/job/ci-platform-meta-agl-demo-verify-CIBT-qemux86-64/283/console * https://jira.automotivelinux.org/browse/SPEC-1209 * reproduce: * build for qemux86-64 and run with runqemu A start job is running for Run pending postinsts (5min 5s / no limit)[K[ [0;31m*[0;1;31m*[0m] A start job is running for Run pending postinsts (5min 5s / no limit)[K[ [0;31m*[0m] A start job is running for Run pending postinsts (5min 6s / no limit)[K[ [0;31m*[0;1;31m*[0m] A start job is running for Run pending postinsts (5min 6s / no limit)[K[ [0;31m*[0;1;31m [0m[0;31m*[0m] A start job is running for Run pending postinsts (5min 7s / no limit)[K[ [0;31m*[0;1;31m*[0m[0;31m* [0m] A start job is running for Run pending postinsts (5min 7s / no limit)[ * Update 2018-02-13: kvm mitigates the issue, will be enabled in the labs * Lab-in-a-box * Training session during AMM * New: * ... ===Next meeting(s)=== Need to skip call due to Embedded World. Next call in 3 & 5 weeks (march 6th and march 20th) same place, same time (13:00 UTC) ==== Meeting 2018/01/30 ==== ===Attendees=== Jan-Simon Möller, Kotaro Hasioto, Liu Wenlong, Tetsuji Yamamoto, Frederic Marec, Ronan Le Martret ===Topics=== * Status updates: * KernelCI * https termination issue being worked on (mixed https/http connection otherwise) * https://github.com/lucj/kernelci-docker/issues/13#issuecomment-360921103 * CIBT / LAVA * need helping hand debugging qemux86-64 target * https://build.automotivelinux.org/job/ci-platform-meta-agl-demo-verify-CIBT-qemux86-64/283/console * https://jira.automotivelinux.org/browse/SPEC-1209 * reproduce: * build for qemux86-64 and run with runqemu A start job is running for Run pending postinsts (5min 5s / no limit)[K[ [0;31m*[0;1;31m*[0m] A start job is running for Run pending postinsts (5min 5s / no limit)[K[ [0;31m*[0m] A start job is running for Run pending postinsts (5min 6s / no limit)[K[ [0;31m*[0;1;31m*[0m] A start job is running for Run pending postinsts (5min 6s / no limit)[K[ [0;31m*[0;1;31m [0m[0;31m*[0m] A start job is running for Run pending postinsts (5min 7s / no limit)[K[ [0;31m*[0;1;31m*[0m[0;31m* [0m] A start job is running for Run pending postinsts (5min 7s / no limit)[ * Fuego * no update * Lab-in-a-box * no update * New topics: * ==== Meeting 2018/01/02 ==== ===Attendees=== Jan-Simon Möller, Frederic Marec, Liu Wenlong, Kyohhei Oki, Ronan Le Martret, Tetsuji Yamamoto ===Topics=== == AGL CIAT meeting == * Status updates: * KernelCI * Setup WIP with LF-IT * https://kernelci-docker.automotivelinux.org:8080/ * Update 2018/01/16: instance up, being moved to production URL * CIBT / LAVA * need helping hand debugging qemux86-64 target * https://build.automotivelinux.org/job/ci-platform-meta-agl-demo-verify-CIBT-qemux86-64/283/console * https://jira.automotivelinux.org/browse/SPEC-1209 * reproduce: * build for qemux86-64 and run with runqemu running for Run pending postinsts (5min 5s / no limit)[K[ [0;31m*[0;1;31m*[0m] A start job is running for Run pending postinsts (5min 5s / no limit)[K[ [0;31m*[0m] A start job is running for Run pending postinsts (5min 6s / no limit)[K[ [0;31m*[0;1;31m*[0m] A start job is running for Run pending postinsts (5min 6s / no limit)[K[ [0;31m*[0;1;31m*[0m[0;31m*[0m] A start job is running for Run pending postinsts (5min 7s / no limit)[K[ [0;31m*[0;1;31m*[0m[0;31m* [0m] A start job is running for Run pending postinsts (5min 7s / no limit)[ * Roadmap for 2018 * From CIAT EG * Lab in a box * Board management * Board on developer's desk * Improved publishing of test results * UI improvements to KCi * Additional tests * Applications * test output should use TAP as output format * work out example test for one application * Platform * ptests (embedded in the yocto recipes) * BSPs * KSelfTest * Fuego integration * Support to run tests as lava jobs (from within Fuego) * New: ===Next meeting(s)=== Call in 2 weeks same place, same time (13:00 UTC) === -------- ==== Meeting 2018/01/02 ==== ===Attendees=== Jan-Simon Möller, Loys Ollivier, Frederic Marec, Liu Wenlong, Ronan ===Topics=== == AGL CIAT meeting == * Open Round * Jan-Simon * Loys * Status updates: * KernelCI * Setup WIP with LF-IT * Need newer docker on CentOS 7 (due to docker compose) * https://jira.automotivelinux.org/browse/SPEC-1208 * Update: new version in git * fixes seed.sh (dropping --gzip) * adds log persistency * https://kernelci-docker.automotivelinux.org:8080/ * CIBT / LAVA * need helping hand debugging qemux86-64 target * https://build.automotivelinux.org/job/ci-platform-meta-agl-demo-verify-CIBT-qemux86-64/283/console * reproduce: * build for qemux86-64 and run with runqemu running for Run pending postinsts (5min 5s / no limit)[K[ [0;31m*[0;1;31m*[0m] A start job is running for Run pending postinsts (5min 5s / no limit)[K[ [0;31m*[0m] A start job is running for Run pending postinsts (5min 6s / no limit)[K[ [0;31m*[0;1;31m*[0m] A start job is running for Run pending postinsts (5min 6s / no limit)[K[ [0;31m*[0;1;31m*[0m[0;31m*[0m] A start job is running for Run pending postinsts (5min 7s / no limit)[K[ [0;31m*[0;1;31m*[0m[0;31m* [0m] A start job is running for Run pending postinsts (5min 7s / no limit)[ ===Next meeting(s)=== Call in 2 weeks same place, same time (13:00 UTC) . Jan 16th 2018. ==== Meeting 2017/12/19 ==== ===Attendees=== Jan-Simon Möller, Tim Bird, Frederic Marec, Liu Wenlong, Tetsuji Yamamoto, Kotaro Hashimoto, Stephane Desneux ===Topics=== == Fuego == == Hosting FUEGO call today - welcome ! == * Release status * 1.2.1 release was made and announced * Fixups to OpenSSL, LTP, and other tests * added support for invoking individual phases * can do 'build' separate from 'run', for any test * next Major release areas of focus * Documentation conversion to reStructuredText * Documentation for end users (not developers) * How to customize tests (criteria.json) * Fuego automated release testing (Fuego selftest) * Under contract with ProFusion * System provisioning * Build and deploy system to target * Work is currently being done in the 'master' branch * Activity * At Japan Jamboree #63 - Tim gave a status report * Another presentation - Tips and tricks for using Fuego (Feature you didn't know Fuego had) * Fuego Japan Hackathon * See http://fuegotest.org/wiki/Japan_Fuego_Hackathon_2017 * OLD * Want 'test' integration with LAVA, instead of 'transport'-level integration * Board farm standards * provisioning - using libvirt, ttc, labgrid or lava * Fuego conversion of docs to reStructuredText * Conversion from Docbook and wiki Markdown to reStructuredText * Allow use of readthedoc.io for online * Build documentation from repository (move from wiki) * Not funded - requested to go back and see if community can perform this * Fuego release self-testing * Use Fuego to do Fuego CI testing * use SeleniumHQ for web interface testing * use 2-container system (driving container and DUT container, with next release of Fuego) * Funding was approved == AGL == == AGL CIAT meeting == * Status updates: * KernelCI * Setup WIP with LF-IT * Need newer docker on CentOS 7 (due to docker compose) * https://jira.automotivelinux.org/browse/SPEC-1208 * CIBT / LAVA * need helping hand debugging qemux86-64 target * https://build.automotivelinux.org/job/ci-platform-meta-agl-demo-verify-CIBT-qemux86-64/283/console * reproduce: * build for qemux86-64 and run with runqemu running for Run pending postinsts (5min 5s / no limit)[K[ [0;31m*[0;1;31m*[0m] A start job is running for Run pending postinsts (5min 5s / no limit)[K[ [0;31m*[0m] A start job is running for Run pending postinsts (5min 6s / no limit)[K[ [0;31m*[0;1;31m*[0m] A start job is running for Run pending postinsts (5min 6s / no limit)[K[ [0;31m*[0;1;31m*[0m[0;31m*[0m] A start job is running for Run pending postinsts (5min 7s / no limit)[K[ [0;31m*[0;1;31m*[0m[0;31m* [0m] A start job is running for Run pending postinsts (5min 7s / no limit)[ * container build * Fixed. * Reports for rc's * https://wiki.automotivelinux.org/agl-distro/releases/4.99.3/test/summary ===Next meeting(s)=== Call in 2 weeks same place, same time (13:00 UTC) . Jan 2nd 2018. ==== Meeting 2017/11/21 ==== ==Attendees== Jan-Simon Möller, Frederic Marec, Daniel Sangorrin, Yuichi Kusakabe, Sebastien Douheret, Liu Wenlong, Tetsuji Yamamoto, Stephane Desneux Ronan, Qiutingting, Kyohhei Oki, Loys Ollivier ==Topics== = AGL = * Status updates: * KernelCI * Loys: * migrated boards to submit to KCi via callback * https://kernelci.automotivelinux.org:4443 * releng scripts used by jenkins submit jobs with proper callbacks now * Next: adding more tests + bugfixes * LAVA * Jan-Simon: * Atm power outage in the core lab * 2017.10 * Jenkins * Had the merge job being retriggered over and over. Fixed now. * Need to more app repos to ci-management/jjb/ci-apps (EOW) * Next release: * rc3 cut-off is 24th Nov * ptest * rc4 cut-off is 8th Dec * ptest * Open Issues: * Kingfisher * aglsetup machine template ? really needed ? * difference is only: kernel builds another dtb in addition * u-boot (usually not used) * additional packages pulled-in (bt firmware, ...) * -> we don't need a separate template * challenge: weston.ini ? runtime detect of KF ? * Preference 1: keep same MACHINE (aglsetup + YP) * Preference 2: keep same YP machine (aglsetup m3ulcb-kf + YP m3ulcb) * Avoid: new MACHINE in YP ! * container build * https://build.automotivelinux.org/view/jjb-housekeeping-jobs/job/snapshot-jjb-container-weekly/65/console * SPEC-979 ==Next meeting(s)== Call in 2 weeks same place, same time (13:00 UTC) . ==== Meeting 2017/11/07 ==== ==Attendees== Jan-Simon, Tim, Daniel, Stephane ==Topics== == Hosting FUEGO call today - welcome ! First 30 minutes Fuego, 2nd 30 minutes AGL-CIAT == = Fuego = * OLD * Fuego status: * 1.2 release was made, but without announcement * Major features: * unified output format (run.json - somewhat compatible with kernelci) * test dependency system (NEED_xxx variables) * complex pass criteria handling (criteria.json) * chart_config.json (foundation for customizing results visualization) * test improvements (e.g. aarch64 support for old tests) * documentation fixups * Pending items for announcements: * LTP parsing of POSIX results * More Documentation fixups * Am now planning on a 1.2.1 release which includes recent bugfixes from Tim, Wenlong and Daniel * NEW: * Good discussion with Jan-Simon and Kevin Hilman about LAVA integration * Currently, we have "transport"-level integration * We need "test"-level integration * Kevin offered to be the contractor for this * Board farm/DUT control automation discussed at ELCE * Decided to use automated-testing@yoctoproject.org for list * https://lists.yoctoproject.org/listinfo/automated-testing * Lots of people doing DUT control projects (libvirt, ttc, labgrid, pduclient) * would be nice to colaborate * would be nice to have a DUT control connection standard * CELP project funding * Discussed 4 Fuego projects: * LAVA integration * needs additional research, CELP deferred funding decision * Renesas wants to see justification for expense, may push funding request through AGL * Fuego size test * also needs additional research - not much interest * Fuego conversion of docs to reStructuredText * Conversion from Docbook and wiki Markdown to reStructuredText * Allow use of readthedoc.io for online * Build documentation from repository (move from wiki) * Not funded - requested to go back and see if community can perform this * Fuego release self-testing * Use Fuego to do Fuego CI testing * use SeleniumHQ for web interface testing * use 2-container system (driving container and DUT container, with next release of Fuego) * Funding was approved * Work to be done by ProFusion * Maybe add support for Jenkins pipelines? = AGL = * Status updates: * KernelCI : test instance up https://kernelci.automotivelinux.org:4443 * EE 5.0RC2 being built , tests will get notifiers to kCI * F2F during AMM minutes ... still WIP, will send email * Loys: what has been done on the kernelCI / releng-scripts / ptest front * ptest wrapper for lava and submit to kernelCI * http://kernelci.dev.baylibre.com/test/board/qemu/job/AGL-kernel-tree/kernel/PTEST-02/ * releng-scripts can now generate jobs for m3, qemu, rpi3 * goal all CI boards should submit through releng-scripts * including notifiers for kernelCI * AMM: * 2 Talk * Lab-in-a-box * Writing test * Workshop on Friday * more tests (help welcome!): * evaluate SPEC-272 and extend services being checked * need to differ between agl-image-ivi (no gfx) and agl-demo-platform (=profiles in the end) * e.g. check appfw with curl call * e.g. check kernel modules being loaded (as per our bbappends+CONFIG options) * can this be a macro or ptest ? * AMM: discuss with Jose how to appfw / application (self-)tests * Next * Discussion about time for call. Alternate time ? * E.g. stick with 13:00 UTC * Alternative time would be e.g. 16:00 UTC * Stick with 13:00 UTC ==Next meeting(s)== Call in 2 weeks same place, same time (13:00 UTC) . ==== Meeting 2017/10/10 ==== ==Attendees== Jan-Simon, Frederic Marec (iot.bzh), Liu Wenlong, Ronan Le Martret, Song Cai, Agustin (Codethink) ==Topics== = AGL = * Status updates: * KernelCI : Setup progressing, IT working on storage backend. * Pyro: * Master now stable and we're about to do an RC1, porter still has issues with gfx. * SPEC-544: FIXED ! Thanks to Scott Murray, we have a patch to mitigate the problem. * CI confirms that the boot now works * more tests (help welcome!): * evaluate SPEC-272 and extend services being checked * need to differ between agl-image-ivi (no gfx) and agl-demo-platform (=profiles in the end) * e.g. check appfw with curl call * e.g. check kernel modules being loaded (as per our bbappends+CONFIG options) * can this be a macro or ptest ? * mid-sized task: can ptest's be wrapped (SPEC-758) * Next * EE 5.0RC1 , tests in kCI * F2F breakout during AMM next tuesday (actually the day before AMM starts, 2pm Salon Mainz-Frankfurt) * Add your name here: https://wiki.automotivelinux.org/eg-ciat/meetings/2017-oct-f2f * another short session during the SAT meeting (friday) * https://wiki.automotivelinux.org/agl-distro/oct2017-f2f ==Next meeting(s)== Call in 4 weeks same place, same time. (Due to ELCE in Prague in 2 weeks) ==== Meeting 2017/09/26 ==== ==Attendees== Jan-Simon, Tim, Stephane, ==Topics== = AGL CIAT = * Status updates: * KernelCI instance in preparation with LFIT (ETA Thu) * Pyro: * Master works in CI, Tests fail more frequently due to SPEC-544! * Dab (morty) working perfectly fine * Test job for EE rc1 (check for the running/required services, SPEC-272, merged, Loys (baylibre) * Document how to add tests and use the generator/templates (c11029/Baylibre) * Open Round: * build chrome as app * igalia working on chrom+wayland (renesas platform) * currently being done as 'platform build' (aka with bitbak/yocto) * issues with ivi-shell, thus fallback to desktop-shell * iot.bzh did research on app build: * https://github.com/iotbzh/chromium-agl-app * build as app can save resources - thus interesting for CI * focus on renesas gen3, qemu broken (libav) * AI: add https://github.com/iotbzh/chromium-agl-app/blob/master/meta-sdk-chrom/recipes-platform/images/agl-demo-platform-crosssdk.bbappend to agl-demo-platform-crosssdk.bb * AI: add JIRA for vetting of other TOOLCHAIN_HOST_TASK TOOLCHAIN_TARGET_TASK for EE. * App builds in CI are for master/snapshot. AI: add release builds for dab but this requires dab branch in app repos (-> devcall/sat) * Next: * kCI token setup for lava.automotivelinux.org * SPEC-544 (serial/udev issue) (Scott/Konsulko) * more tests (help welcome!): * evaluate SPEC-272 and extend services being checked * need to differ between agl-image-ivi (no gfx) and agl-demo-platform (=profiles in the end) * e.g. check appfw with curl call * e.g. check kernel modules being loaded (as per our bbappends+CONFIG options) * can this be a macro or ptest ? * mid-sized task: can ptest's be wrapped (SPEC-) * AI: test latest Fuego with nbd patchset ==Next meeting(s)== In 2 weeks ==== Meeting 2017/09/12 ==== ==Attendees== Jan-Simon, Tim Bird, Agustin, Daniel, Liu Wenglong, Patrick Titiano, Hashimoto-san, Jerome Brunet ==Topics== == Hosting FUEGO call today - welcome ! == * NEW: * Fuego status: * Working towards 1.2 release * 4 main items (see http://bird.org/fuego-1.2/Releases and http://bird.org/fuego-1.2/Release_1.2_To_Do) * unified output format (run.json) - basically done * AGL/JTA reports - in progress (see html table work by Daniel) * Full LTP support - basically done, needs testing * Wenlong Liu and Cai Song have been testing, and submitted patches (thanks!) * updated wiki documentation - started * long-term plan: move to reStructuredText (readthedocs/sphynx) * LAVA integration - was deferred (had added target_setup_link/target_teardown_link) * Jan said I can integrate the existing pull request * Tim will do that * Broken stuff * Fuego plots on Jenkins pages * put chart processing in the parser, not mod.js (part of flot package) * output a flot-friendly json file (flot_chart_data.json) * chart_config.json * Fuego standalone plots * New Features * ftc gen-report (prototype, output=text for now) * criteria.json * testplan_ltsi.json * html tables for functional tests on Jenkins * results.json * new presentation: Integration of Fuego with Gitlab * plans for the future: * move documentation into Fuego repository (under git control) * have ftc gen-report output multiple formats * text, html, pdf, xunit, TAP) * year 2038 test * big push for new tests * kselftest * upstream kernel is converting to TAP output * CAN bus test * integrate Khiem Nguyen's test (LTSI/Renesas board test) * CIP tests * AGL tests * BayLibre - doing some test development for AGL (infrastructure for kernelci) * Also doing a test-in-a-box * Tim doing Linaro Connect presentation on test standards * pduclient * tguids * run.json (kerneci results format - with fuego stuff) == AGL CIAT == * Status updates: * KernelCI additions in mainline now for displaying testresults (not just boot) * Instance being setup by LFIT * Pyro: * Master broken in CI due to failing builds * Dab (morty) working perfectly fine * Next: * Planned for EOW next week KernelCI instance (LFIT) * Test job for EE rc1 (check for the running/required services jsmoelelr) * SPEC-544 (serial/udev issue) (Scott/Konsulko) * Document how to add tests and use the generator/templates (Kevin/Baylibre) ==Next meeting(s)== In 2 weeks (2017/09/26) ==== Meeting 2017/08/29 ==== ==Attendees== Jan-Simon, Agustin, Daniel, Hashimoto-san, Liu Wenglong, Patrick Titiano, Robert Marshall ==Topics== * Board@Desk presentation (Agustin) * Slides can be found in the [[https://wiki.linuxfoundation.org/_media/civilinfrastructureplatform/testing_at_cip__b_d.pdf|CIP wiki in .pdf]] format. * Open round ==Next meeting(s)== In 2 weeks -------- ==== Meeting 2017/08/14 ==== ==Attendees== ==Topics== == Hosting FUEGO call today - welcome ! == * NEW: * We'll skip today's Fuego call then. * The next CIAT meeting is in 2 weeks. ==Next meeting(s)== In 2 weeks -------- ==== Meeting 2017/08/01 ==== ==Attendees== Jan-Simon, Walt, Cai, Jerome, Keisuke Iwata, Liu Wenlong, Agustin ==Topics== * OLD topics: * Hardware test template for qemu (x86-64) * **Update 01/08**: Enabled * planning start the 'generic-*' * Service/Unit 'up' test * **Update 01/08**: * * Lava upstream work: * **Update 01/08**: * Merged: https://review.linaro.org/#/c/20688/ * WIP: https://review.linaro.org/#/c/17088/ * KernelCI work: * **Update 01/08**: * https://jira.automotivelinux.org/browse/SPEC-746 * Multi-lava-lab support in CI merged * Focus: test submission template, likely jinja2 templates as in lava * https://github.com/khilman/lava-ci-staging/commit/3da1fbf931057d55f46c46194ffdf433a9b0e881 * http://lava.baylibre.com:10080/scheduler/job/923/definition * JIRA review: * Reviewing https://jira.automotivelinux.org/issues/?filter=10405 ==Next meeting(s)== In 2 weeks Aug 15th, 12:00 UTC (joint AGL + FUEGO) ==== Meeting 2017/07/18 ==== ==Attendees== Jan-Simon Möller, Cai Song, Keisuke Iwata, Liu Wenlong, Stephane Desneux, Daniel Sangorrin ==Topics== == First ~30 minutes - hosting FUEGO call today - welcome ! == * NEW: * Top issues: - unified json output - lava integration - Daniel provided kernel build test - flot plugin, output refactoring is in support some new test comparison operation and reporting * Major feature remaining for 1.2 is the unified json output, and support for AGL reports - Goal would be to release 1.2 in mid-July. * Unordered List Item* Unified Output Format (UOF) * Daniel showed format of reference.json, with new "pass-criteria" * Unordered List Item Tim talked about needing to share pass-criteria, and how to store and use them for a test * specifically alternate pass-criteria * Daniel showed different files for the new output format * schema, reference.json, run.json, results.json, and png of possible report format * some questions came up: * should the results.json (the aggregate data from multiple runs) contain the pass-criteria or reference data)? * it's easier to parse for the plotter without this * should pass-criteria be separate from the reference.json file? * should customized or alternate pass-criteria be stored in diff format or as full json files? * code to produce the html is in Daniel's next repository * Daniel has a new interface for generating the measurement data (parser.py) * parser.py builds a data structure, and calls plib.process_data to both record the data (measurements) and evaluate the pass criteria from the reference.json file * Tim commented that the bonnie parser.py could be simplified with a function to put operations together * something like: add_measure("block.read","speed",float(5)) * More discussion of the issues will be raised on the mailing lists * interesting option is to use a web-based transfer to target * Tim started a prototype of this * Could transfer commands as well as test materials, using a client pull rather than host push * need an agent on target for this * eliminates the need for an ssh connection, but requires a web host (maybe can use Jenkins for this) * OLD - some miscellaneous technical issues * automatically create a view for each board (using regex of: ".*") (done) * put log links (and others) in description even if job fails (done) * don't require buildsetter plugin (defer) * automatically detect log, plot and xlsx files * support partial results * lots of times there is a test log there, even if the build failed * try to eliminate need for extralinks in testplan file (done - not there anymore) * eliminate need for timeout in testplan file (not needed, if default is used) * do a better job of automatically calculating timeouts * this is dependent on test, spec, board and other parameter * NEW: * a * b == AGL-CIAT Meeting: == * OLD topics: * Related issue: device_type names in lava need to be converted to match DTB name * **Update 04/07**: In preparation * **Update 18/07**: ETA this week to cut-over all labs * Stephane: weston startup to be worked on. Also looks into virtual screens for simulation purposes (e.g. including faking EDIDs). Tests must be independent of the compositor going forward. * Tests initially +/- or Fail/Pass * testdefinition: https://git.automotivelinux.org/src/qa-testdefinitions/tree/test-suites/short-smoke * test run: https://lava.automotivelinux.org/scheduler/job/1798 * Next: enable on all boards and activate in ci * **Update 04/07**: busy w/ dd prep * **Update 18/07**: not yet * SPEC-495 - Template for apps (cmake, tbd. qmake, fallback Makefile) * Autobuild script for CI: * qmake needed for current demo apps * requirements: git clone (or git submodule) , clone and calling of script will build app with autodetected method (qmake, cmake, autotools, Makefile) * app templates usable as git submodules (atm: cmake) * Workflow: * clone (recursive) * [[ -f conf.d/autobuild/agl/autobuild ]] && conf.d/autobuild/agl/autobuild package DEST=$(pwd)/package/ (pickup is package/*.wgt) * + generic procedure (call make as fallback) * new JIRA for qmake macros * **Update 20/06**: Templates enhanced, no Autobuild for CI, yet * **Update 18/07**: Changesets merged in gerrit, doc pending, qmake pending * JIRA review: * Reviewing https://jira.automotivelinux.org/issues/?filter=10405 * NEW topics: * a ==Next meeting(s)== In 2 weeks Aug 1st, 12:00 UTC ---- ==== Meeting 2017/07/04 ==== ==Attendees== Jan-Simon, Daniel, Iwata-san, Hashimoto-san, Liu Wenlong, Oki-san, Stephane, Agustin ==Topics== * OLD topics: * Hardware test template for gen3 * **Update 04/07**: Added, activation changeset pendinng in gerrit * Related issue: device_type names in lava need to be converted to match DTB name * **Update 04/07**: In preparation * Stephane: weston startup to be worked on. Also looks into virtual screens for simulation purposes (e.g. including faking EDIDs). Tests must be independent of the compositor going forward. * Tests initially +/- or Fail/Pass * Next: create check if services are up * https://jira.automotivelinux.org/browse/SPEC-275 * https://jira.automotivelinux.org/browse/SPEC-305 * Target have tests scripts * Build-up: * basic "service up" * then "client connect" * then "request/answer" * **Update 20/06**: WIP * testdefinition: https://git.automotivelinux.org/src/qa-testdefinitions/tree/test-suites/short-smoke * test run: https://lava.automotivelinux.org/scheduler/job/1798 * Next: enable on all boards and activate in ci * **Update 04/07**: busy w/ dd prep * SPEC-495 - Template for apps (cmake, tbd. qmake, fallback Makefile) * add optional script there, bbclass add generic script if no script is there. * yocto SDK installs too-much debug-packages (e.g. from platform) * https://git.automotivelinux.org/apps/app-templates/tree/ * templates to be used with submodules * cmake template and sample https://github.com/iotbzh/helloworld-service * Autobuild script for CI: * qmake needed for current demo apps * requirements: git clone (or git submodule) , clone and calling of script will build app with autodetected method (qmake, cmake, autotools, Makefile) * **Update 20/06**: Templates enhanced, no Autobuild for CI, yet * app templates usable as git submodules (atm: cmake) * Workflow: * clone (recursive) * [[ -f conf.d/autobuild/agl/autobuild ]] && conf.d/autobuild/agl/autobuild package DEST=$(pwd)/package/ (pickup is package/*.wgt) * + generic procedure (call make as fallback) * new JIRA for qmake macros * NEW topics: * Reviewing https://jira.automotivelinux.org/issues/?filter=10405 * https://jira.automotivelinux.org/browse/SPEC-502 ?? Review later * https://jira.automotivelinux.org/browse/SPEC-237 ?? Status ? ==Next meeting(s)== In 2 weeks July 18th, 12:00 UTC (joint AGL + FUEGO) ---- ==== Meeting 2017/06/20 ==== ==Attendees== ==Topics== == First ~30 minutes - hosting FUEGO call today - welcome ! == * OLD: * a * NEW: * b == 2nd ~30 minutes - hosting AGL CIAT call today - welcome ! == * OLD topics: * Hardware test template for gen3 * **Update 13/06**: Activation in CI pending * **Update 20/06**: no time, yet * Related issue: device_type names in lava need to be converted to match DTB name * Stephane: weston startup to be worked on. Also looks into virtual screens for simulation purposes (e.g. including faking EDIDs). Tests must be independent of the compositor going forward. * Tests initially +/- or Fail/Pass * Next: create check if services are up * https://jira.automotivelinux.org/browse/SPEC-275 * https://jira.automotivelinux.org/browse/SPEC-305 * Target have tests scripts * Build-up: * basic "service up" * then "client connect" * then "request/answer" * **Update 13/06**: Examples are here: * **Update 20/06**: WIP * testdefinition: https://git.automotivelinux.org/src/qa-testdefinitions/tree/test-suites/short-smoke * test run: https://lava.automotivelinux.org/scheduler/job/1798 * Next: enable on all boards and activate in ci * SPEC-495 - Template for apps (cmake, tbd. qmake, fallback Makefile) * Move new templates to gerrit repo (SPEC-556) * add optional script there, bbclass add generic script if no script is there. * yocto SDK installs too-much debug-packages (e.g. from platform) * https://git.automotivelinux.org/apps/app-templates/tree/ * templates to be used with submodules * cmake template and sample https://github.com/iotbzh/helloworld-service * Autobuild script for CI: * qmake needed for current demo apps * requirements: git clone (or git submodule) , clone and calling of script will build app with autodetected method (qmake, cmake, autotools, Makefile) * **Update 20/06**: Templates enhanced, no Autobuild for CI, yet * NEW topics: * Reviewing https://jira.automotivelinux.org/issues/?filter=10405 * https://jira.automotivelinux.org/browse/SPEC-502 ?? Review later * https://jira.automotivelinux.org/browse/SPEC-237 ?? Status ? ==Next meeting(s)== In 2 weeks ... but Independency Day in US July 4th, 12:00 UTC -------- ==== Meeting 2017/06/13 ==== ==Attendees== Jan-Simon, Walt, Michael, Agustin, Cai Song, Daniel, Liu Wenlong, Stephane Desneux, Matsuzawa-san, Ronan Le Martret, Kyohhei Oki, ==Topics== * OLD topics: * Hardware test template for gen3 * **Update 13/06**: Solved, board online. * **Update 13/06**: Activation in CI pending * lava-test-shell (overlay), LAVA-911, SPEC-555 * **Update 13/06**: works with transfer_overlay and webserver on slave. * Stephane: weston startup to be worked on. Also looks into virtual screens for simulation purposes (e.g. including faking EDIDs). Tests must be independent of the compositor going forward. * Tests initially +/- or Fail/Pass * Next: create check if services are up * https://jira.automotivelinux.org/browse/SPEC-275 * https://jira.automotivelinux.org/browse/SPEC-305 * Target have tests scripts * Build-up: * basic "service up" * then "client connect" * then "request/answer" * Update 23/95: no update * Update 13/06: script ? * **Update 13/06**: Examples are here: * testdefinition: https://git.automotivelinux.org/src/qa-testdefinitions/tree/test-suites/short-smoke * test run: https://lava.automotivelinux.org/scheduler/job/1798 * Next: enable on all boards and activate in ci * SPEC-495 - Template for apps (cmake, tbd. qmake, fallback Makefile) * Move new templates to gerrit repo (SPEC-556) * add optional script there, bbclass add generic script if no script is there. * yocto SDK installs too-much debug-packages (e.g. from platform) * https://git.automotivelinux.org/apps/app-templates/tree/ * templates to be used with submodules * cmake template and sample https://github.com/iotbzh/helloworld-service * Autobuild script for CI: * qmake needed for current demo apps * requirements: git clone (or git submodule) , clone and calling of script will build app with autodetected method (qmake, cmake, autotools, Makefile) * SDK and qt headers not included * Fixed sdk build here: * arm32 : https://download.automotivelinux.org/AGL/snapshots/master/latest/raspberrypi3/deploy/sdk/ * aarch64 : https://download.automotivelinux.org/AGL/snapshots/master/latest/dragonboard-410c/deploy/sdk/ * x86-64 : https://download.automotivelinux.org/AGL/snapshots/master/latest/intel-corei7-64/deploy/sdk/ * can be used for targets that have the same architecture (e.g. rpi3 also for porter or vayu, dragon for m3/h3, intel-corei7-64 for qemux86-64) * **Update 13/06**: App builds active in jenkins. currently using above URL's (not yet the container) ==Next meeting(s)== next week (+ fuego) June 20, 12:00 UTC -------- ==== Meeting 2017/05/23 ==== ==Attendees== Jan-Simon, Walt Miner, Harunobu Kurokawa, Liu Wenlong, Patrick Titiano, Takahi Matsuzawa, Tim Bird, Cai Song, Kyohhei Oki , Agustin Benito Bethencourt ==Topics== == First ~30 minutes - hosting FUEGO call today - welcome ! == * OLD: * Top issues: - unified json output - patch for functional tests produce a results.json file - discussion on the kernelci results schema (on the fuego mailing list) - Kevin Hillman started discussion - Daniel responded - Daniel added some fields to the run.json (don't know if that's related to discussion) - Lots of work already done in kernelci - should reuse their expertise (if in doubt) - Will probably err on the side of too much info - contact xxx for discussion of fields and schema used by kernelci - Song said that they are working on LTP as higher priority - lava integration - ov_transport_connect/disconnect is goal - have added TARGET_TEARDOWN_LINK - shorter connection times - refactored connection/disconnection location in pre_test and post_test - standalone build phase - prototype with LTP (build_only) - specialized solution - would like a more general one - maybe something like rpm phase selection - Would like to execute jobs in batch job in the same order as the testplan - this would support having a "board setup" job at the beginning of a sequence of jobs - unify the test types - gotten rid of distinction between functional and benchmark tests - main.sh - every test, whether Functional or Benchmark, now uses 'fuego_test.sh' as the base script name. - Daniel provided kernel build test - added support for git repository checkout and build - Has a patch in-flight that does kselftest - flot plugin, output refactoring is in support some new test comparison operation and reporting - goal is to support previous JTA use cases and ALG-JTA use cases for reporting in 1.2 release - Testing BOF at ALS, June 2, 2:00 pm - Tentative plan private meeting at ALS, on May 31, at 10:00 am. - Major feature remaining for 1.2 is the unified json output, and support for AGL reports - Goal would be to release 1.2 in mid-July. * status of toshiba/sony merge * made the 1.1 release ("blaze") * resolving technical issues with Tim's 'next' branch (based on Daniel's 'next' branch) (done) * priority list for AGL, to start that merge * LAVA integration * is 'uniform test output' a high-priority feature? (yes) * interesting option is to use a web-based transfer to target * Tim started a prototype of this * Could transfer commands as well as test materials, using a client pull rather than host push * need an agent on target for this * eliminates the need for an ssh connection, but requires a web host (maybe can use Jenkins for this) * planning for an official fuego release (done) * 1.1.0 release, code name: blaze * some miscellaneous technical issues * TESTNAME (base script name) is derived from the test name (.testname) * use of spec name in job name, instead of plan name (done) * add node and job scripts to ftc (done) * eliminate fuego-ro/scripts (done) * move /fuego-ro/conf/boards to /fuego-ro/boards (done) * AGL has lava scripts in /fuego-ro, and some board-related stuff * in boards, auth files for LAVA * in fuego-ro/scripts - 2 scripts (for link setup and teardown) * defer looking at this right now * automatically create a view for each board (using regex of: ".*") (done) * put log links (and others) in description even if job fails (done) * don't require buildsetter plugin (defer) * automatically detect log, plot and xlsx files * support partial results * lots of times there is a test log there, even if the build failed * try to eliminate need for extralinks in testplan file (done - not there anymore) * eliminate need for timeout in testplan file (not needed, if default is used) * do a better job of automatically calculating timeouts * this is dependent on test, spec, board and other parameter * NEW: == AGL-CIAT Meeting: == * OLD topics: * Hardware test template for porter & vayu * Update 05/23: both active * issue with serial port bringup/systemd - CIP had similar issue * See https://jira.automotivelinux.org/browse/SPEC-544 * Leads to false positives * gen3 poweron issue - WOL capabilities ? Likely reworking board. * One missing piece: lava-test-shell (overlay), LAVA-911, SPEC-555 * Stephane: weston startup to be worked on. Also looks into virtual screens for simulation purposes (e.g. including faking EDIDs). Tests must be independent of the compositor going forward. * Tests initially +/- or Fail/Pass * Next: create check it is up * https://jira.automotivelinux.org/browse/SPEC-275 * https://jira.automotivelinux.org/browse/SPEC-305 * Target have tests scripts * Build-up: * basic "service up" * then "client connect" * then "request/answer" * Update 05/23: no update * SPEC-495 - Script to build apps * Move new templates to gerrit repo (SPEC-556) * add optional script there, bbclass add generic script if no script is there. * yocto SDK installs too-much debug-packages (e.g. from platform) * NEW topics: * SDK and qt headers not included * Fixed sdk build here: * arm32 : https://download.automotivelinux.org/AGL/snapshots/master/latest/raspberrypi3/deploy/sdk/ * aarch64 : https://download.automotivelinux.org/AGL/snapshots/master/latest/dragonboard-410c/deploy/sdk/ * x86-64 : https://download.automotivelinux.org/AGL/snapshots/master/latest/intel-corei7-64/deploy/sdk/ * can be used for targets that have the same architecture (e.g. rpi3 also for porter or vayu, dragon for m3/h3, intel-corei7-64 for qemux86-64) * WIP: application builds in jenkins ==Next meeting(s)== June 6, 12:00 UTC -------- ==== Meeting 2017/04/25 ==== ==Attendees== Jan-Simon, Walt, Michael Fabry, Ronan Le Martret, Stephane Desneux, Agustin ==Topics== * OLD topics: * Hardware test template for porter & vayu * ** UPDATE 04/11 **: jsmoeller working on porter, bootloader solved, working on userspace. * ** UPDATE 04/25 **: No time. * ** UPDATE 05/02 **: porter merged c8635, vayu in review 9295 * rpi3, porter, vayu (in review) active * issue with serial port bringup/systemd - CIP had similar issue. * gen3 poweron issue - WOL capabilities ? * a) automate button-press * b) add-on board (renesas add-on ?? or custom ??) * c) rework the board (SMD soldering) * One missing piece: lava-test-shell (overlay), LAVA-911, SPEC-555 * Stephane: weston startup to be worked on. Also looks into virtual screens for simulation purposes (e.g. including faking EDIDs). Tests must be independent of the compositor going forward. * Tests initially +/- or Fail/Pass * Next: create check it is up * https://jira.automotivelinux.org/browse/SPEC-275 * https://jira.automotivelinux.org/browse/SPEC-305 * Target have tests scripts * Build-up: * basic "service up" * then "client connect" * then "request/answer" * ** Update 04/11 **: AI (stephane/ronan): test to check weston "service up" (as root) * ** Update 04/25 **: * https://git.automotivelinux.org/AGL-JTA-core/tree/engine/tests/additional/Additional.check_wayland_weston/check_wayland_weston.sh * Need example for yaml in qa-testdefinitions (stress + weston) * AI: create script from above (omit RPM), if it runs on target, we can create a qa-testdefinition template in AGL gerrit (Stephane, Jsmoeller) * **Update 05/02 **: no script, yet. qa-testdefinition blocked by test-shell issue above (SPEC-555). * SPEC-495 - Script to build apps * Move new templates to gerrit repo (SPEC-556) * add optional script there, bbclass add generic script if no script is there. * SDK discussion: * need for DEFAULT_TUNE * less sstate-cache reuse than expected, needs debugging * some do "generic-x86-64, generic-arm, generic-aarch64" (w/o kernel/u-boot) * yocto SDK installs too-much debug-packages (e.g. from platform) * NEW topics: * SPEC- ==Next meeting(s)== Time change due to scheduling conflict: May 23, 12:00 UTC -------- ==== Meeting 2017/04/25 ==== ==Attendees== Jan-Simon, Walt, Song Cai, Francis Ielsch, Liu Wenglong, Stephane Desneux, Sebastien Douheret, Agustin, Kusakabe-san ==Topics== == First ~30 minutes - hosting FUEGO call today - welcome ! == * Top issues: - unified json output - patch for functional tests produce a results.json file - discussion on the kernelci results schema (on the fuego mailing list) - Kevin Hillman started discussion - Daniel responded - Daniel added some fields to the run.json (don't know if that's related to discussion) - Lots of work already done in kernelci - should reuse their expertise (if in doubt) - Will probably err on the side of too much info - contact xxx for discussion of fields and schema used by kernelci - Song said that they are working on LTP as higher priority - lava integration - ov_transport_connect/disconnect is goal - have added TARGET_TEARDOWN_LINK - shorter connection times - refactored connection/disconnection location in pre_test and post_test - standalone build phase - prototype with LTP (build_only) - specialized solution - would like a more general one - maybe something like rpm phase selection - unify the test types - gotten rid of distinction between functional and benchmark tests - main.sh - Daniel provided kernel build test - flot plugin, output refactoring is in support some new test comparison operation and reporting - goal is to support previous JTA use cases and ALG-JTA use cases for reporting in 1.2 release * status of toshiba/sony merge * made the 1.1 release ("blaze") * resolving technical issues with Tim's 'next' branch (based on Daniel's 'next' branch) (done) * priority list for AGL, to start that merge * LAVA integration * is 'uniform test output' a high-priority feature? (yes) * interesting option is to use a web-based transfer to target * Tim started a prototype of this * Could transfer commands as well as test materials, using a client pull rather than host push * need an agent on target for this * eliminates the need for an ssh connection, but requires a web host (maybe can use Jenkins for this) * planning for an official fuego release (done) * 1.1.0 release, code name: blaze * some miscellaneous technical issues * TESTNAME (base script name) is derived from the test name (.testname) * use of spec name in job name, instead of plan name (done) * add node and job scripts to ftc (done) * eliminate fuego-ro/scripts (done) * move /fuego-ro/conf/boards to /fuego-ro/boards (done) * AGL has lava scripts in /fuego-ro, and some board-related stuff * in boards, auth files for LAVA * in fuego-ro/scripts - 2 scripts (for link setup and teardown) * defer looking at this right now * automatically create a view for each board (using regex of: ".*") (done) * put log links (and others) in description even if job fails (done) * don't require buildsetter plugin (defer) * automatically detect log, plot and xlsx files * support partial results * lots of times there is a test log there, even if the build failed * try to eliminate need for extralinks in testplan file (done - not there anymore) * eliminate need for timeout in testplan file (not needed, if default is used) * do a better job of automatically calculating timeouts * this is dependent on test, spec, board and other parameter == AGL-CIAT Meeting: == * OLD topics: * Hardware test template for porter & vayu * ** UPDATE 04/11 **: jsmoeller working on porter, bootloader solved, working on userspace. * ** UPDATE 04/25 **: No time. * Stephane: weston startup to be worked on. Also looks into virtual screens for simulation purposes (e.g. including faking EDIDs). Tests must be independent of the compositor going forward. * Tests initially +/- or Fail/Pass * Next: create check it is up * https://jira.automotivelinux.org/browse/SPEC-275 * https://jira.automotivelinux.org/browse/SPEC-305 * Target have tests scripts * Build-up: * basic "service up" * then "client connect" * then "request/answer" * ** Update 04/11 **: AI (stephane/ronan): test to check weston "service up" (as root) * ** Update 04/25 **: * https://git.automotivelinux.org/AGL-JTA-core/tree/engine/tests/additional/Additional.check_wayland_weston/check_wayland_weston.sh * Need example for yaml in qa-testdefinitions (stress + weston) * AI: create script from above (omit RPM), if it runs on target, we can create a qa-testdefinition template in AGL gerrit (Stephane, Jsmoeller) * SDK discussion: * need for DEFAULT_TUNE * Slides: * https://docs.google.com/presentation/d/1_VSGfheItufCCP_rbGBDg3-2miCyh02-yZva6m8cyUE/edit?usp=sharing * https://jira.automotivelinux.org/browse/SPEC-491 * **Update 04/25**: Likely 4 SDKs (arm32 high, arm32 medium, aarch64, qemux86/core-i7 combined) - PoC under development, input/feedback/patches welcome. * 'default' IMAGE_FSTYPES (not yet started) * https://jira.automotivelinux.org/browse/SPEC-492 * **Update 04/11**: work not started * NEW topics: * SPEC-495 - Script to build apps ==Next meeting(s)== Time change due to scheduling conflict: May 2, 12:00 UTC May 23, 12:00 UTC ==== Meeting 2016/04/11 ==== ==Attendees== * Jan-Simon, Agustin, Tim, Stephane, Song, Liu, Kyohhei, Sebastien, Yasuyuki, Ronan, Kusakabe-san ==Topics== * Roll call * collecting topic additions * Read-out of the F2F * Note: https://docs.google.com/document/d/1lbFeWsxizL-9ZJKOkuwbEjSFptMojqcYXD_5lqAjgFs/edit?usp=sharing * Slides: https://docs.google.com/presentation/d/1fcJTS6XmmjU-de1doefXHh55jNqS9-Dehal7kvDPPoY/edit?usp=sharing * open status round OLD: * Hardware test template for porter & vayu * ** UPDATE 04/11 **: jsmoeller working on porter, bootloader solved, working on userspace. * Stephane: weston startup to be worked on. Also looks into virtual screens for simulation purposes (e.g. including faking EDIDs). Tests must be independent of the compositor going forward. * Tests initially +/- or Fail/Pass * Next: create check it is up * https://jira.automotivelinux.org/browse/SPEC-275 * https://jira.automotivelinux.org/browse/SPEC-305 * Target have tests scripts * Build-up: * basic "service up" * then "client connect" * then "request/answer" * Update 04/11: * AI (stephane/ronan): test to check weston "service up" (as root) * SDK discussion: * need for DEFAULT_TUNE * https://jira.automotivelinux.org/browse/SPEC-491 * **Update 03/28**: jsmoeller working on this topic * found by printing ' bitbake -e agl-demo-platform | grep "^TUNE_FEATURES=" ' * for arm/32bit: lowest common denominator is: * "arm armv7ve vfp thumb neon vfpv4 callconvention-hard" xxxx * ^^^^^ armv7-a with virtualization extensions * other boards w/o virtualization extensions would need (e.g. beaglebone) * "arm armv7e vfp thumb neon vfpv3 callconvention-hard" * for arm/64bit: "aarch64" (no specific core tunings) -> armv7 * Issue it that there are no settings for "neon"/"vfpv4" in the TUNE_FEATURES * Affected components: * Linux: * kernel (?) * Graphics/Video: * gstreamer plugins/codecs * pixman * libjpeg * libpng (! recipe TUNE_FEATURES neon !) <<<<<<<<< * mpg123 * x264 * fftw (Fast Fourier transform) * Audio: * flac * pulseaudio * speexdsp * Engines/interpreters: * nodejs * rust * Reference: http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.den0024a/ch07s03.html * Direction is to actually runtime-detect the flags as e.g. x86 does. * JIRA * We need to plan for 3-4 SDK's (1-2 arm32, 1xaarch64, x86-64) * 'default' IMAGE_FSTYPES (not yet started) * https://jira.automotivelinux.org/browse/SPEC-492 * **Update 04/11**: work not started * NEW topics: * Organization of images ? * ==Next meeting== Apr 25th, same time (Fuego call) ==== Meeting 2017/03/28 ==== ==Attendees== * Jan-Simon Möller, Yasuyuki Matsuura, Ronan Le Martret, Stephane Desneux ==Topics== OLD: * Hardware test template for porter & vayu * ** UPDATE 03/28 **: jsmoeller working on porter, working on issue in bootloader configuration. * Stephane: weston startup to be worked on. Also looks into virtual screens for simulation purposes (e.g. including faking EDIDs). Tests must be independent of the compositor going forward. * Tests initially +/- or Fail/Pass * **Update 03/28**: move to weston 1.11, differences in startup of weston (weston-start) * poky/meta/recipes-graphics/wayland/weston-init/weston-start * https://jira.automotivelinux.org/browse/SPEC-477 * https://gerrit.automotivelinux.org/gerrit/#/c/8895/ * https://gerrit.automotivelinux.org/gerrit/#/c/8897/ * https://gerrit.automotivelinux.org/gerrit/#/c/8921/ * Next: create check it is up * https://jira.automotivelinux.org/browse/SPEC-275 * https://jira.automotivelinux.org/browse/SPEC-305 * Target have tests scripts * Build-up: * basic "service up" * then "client connect" * then "request/answer" * We need to adapt AGL-JTA to the changes introduced by FUEGO's next branch ! * mainly changes in the job templates * Maintainers of AGL-JTA, please step-up ! * Assigned to: ? Cai Song - could you take that on ? * https://jira.automotivelinux.org/browse/SPEC-502 * SDK discussion: * need for DEFAULT_TUNE * https://jira.automotivelinux.org/browse/SPEC-491 * **Update 03/28**: jsmoeller working on this topic * for arm/32bit: lowest common denominator is: "arm armv7ve vfp thumb neon vfpv4 callconvention-hard" * found by printing ' bitbake -e agl-demo-platform | grep "^TUNE_FEATURES=" ' * for arm/64bit: "aarch64" (no specific core tunings) -> armv7 * Issue it that there are no settings for "neon"/"vfpv4" in the TUNE_FEATURES * Affected components: * Linux: * kernel (?) * Graphics/Video: * gstreamer plugins/codecs * pixman * libjpeg * libpng (! recipe TUNE_FEATURES neon !) * mpg123 * x264 * fftw (Fast Fourier transform) * Audio: * flac * pulseaudio * speexdsp * Engines/interpreters: * nodejs * rust * Reference: http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.den0024a/ch07s03.html * Direction is to actually runtime-detect the flags as e.g. x86 does. * JIRA * 'default' IMAGE_FSTYPES (not yet started) * https://jira.automotivelinux.org/browse/SPEC-492 * **Update 03/28**: work not started * Longer test reports need to be uploaded lateron to a frontend (which could be kernelci) * NEW topics: * CFLAGS / SPEC-501 * ==Next meeting== * F2F in Karlsruhe - hangout dial-in tbd * Apr 11th , 12:00 UTC ==== Meeting 2017/03/21 ==== ==Attendees== Jan-Simon, Daniel, Tim, ==Topics== - hosting FUEGO call * Hosting FUEGO call today - welcome * Top issues: - determining the priority list for AGL, to start that merge - planning an official Fuego release * status of toshiba/sony merge - resolving technical difficulties with Daniel's 'next' branch * primarily log paths, and post_test call issues * running scripts as set -e * some ovgen.py features are missing (specifically - ability to run without a default spec) * resolving technical issues with Tim's 'next' branch (based on Daniel's 'next' branch) * log paths - have consolidates logs for a test * Jenkins BUILD_ID changed from timestamp to build number * post_test - need processes to pkill * this is a cleanup operation, and should be in the base script * need test_cleanup * for most tests this could be empty * some tests might call a new function: kill_procs * it calls ov_rootfs_pkill underneath * priority list for AGL, to start that merge * LAVA integration * need to call link_setup very early or very late in test setup * if we call early, then board is reserved during the build, which can time out * maybe switch to not run board setup and board pre_check until after build (reorder these phases) * proposal is to use ov_transport_connect and ov_transport_disconnect * LTP takes a long time to build * want to do a pre_check before the test - may get config information from target before build * is an argument * usually we build with no knowledge of the board * tim's idea is to have a separate test to populate the board (device dictionary) with info about the board * some tests may want to probe during pre_test, and use that to affect the test * if you can set the variables manually, you don't have to run a board_check * one idea is to have configurable phase order (pipeline) * is 'uniform test output' a high-priority feature? * interesting option is to use a web-based transfer to target * planning for an official fuego release * 1.1.0 release, code name: blaze (or brimestone?) - some miscellaneous technical issues * TESTNAME (base script name) is derived from the test name (.testname) * I propose changing this to fuego_test.sh, so they're all the same * any objections - pros?/cons? * pro - can eliminate TESTNAME, and remove from test.yaml file also * pro - less chance of conflict with a script by that name in the test itself * pro - can grep base scripts more easily (all names the same, and no conflicts with other names) * con - can't identify test from it's own name. It's dependent on the test directory to disambiguate it from other tests' base scripts. * con - can't use base-script (arg0?) to find test directory. I'm not sure if this is needed * use of spec name in job name, instead of plan name * this allows creating jobs by spec instead of plan * this supports multiple specs for the same test, in a single plan * add node and job scripts to ftc * eliminate fuego-ro/scripts * move /fuego-ro/conf/boards to /fuego-ro/boards * AGL has lava scripts in /fuego-ro, and some board-related stuff * in boards, auth files for LAVA * in fuego-ro/scripts - 2 scripts (for link setup and teardown) * automatically create a view for each board (using regex of: ".*") * put log links (and others) in description even if job fails * don't require buildsetter plugin * automatically detect log, plot and xlsx files * support partial results * lots of times there is a test log there, even if the build failed * try to eliminate need for extralinks in testplan file * eliminate need for timeout in testplan file * do a better job of automatically calculating timeouts * this is dependent on test, spec, board and other parameter * OLD topics: * We need to adapt AGL-JTA to the changes introduced by fuego's next branch ! * mainly changes in the job templates * Maintainers of AGL-JTA, please step-up ! * Assigned to: * HW tests / CI cycle * SDK discussion: * need for DEFAULTTUNE (not yet started) * https://jira.automotivelinux.org/browse/SPEC-491 * 'default' IMAGE_FSTYPES (not yet started) * https://jira.automotivelinux.org/browse/SPEC-492 * NEW topics: * . * . ==Next meeting== Mar 28th , 12:00 UTC ==== Meeting 2017/02/28 ==== ==Attendees== Jan-Simon Möller, Takahi Matsuzawa, Song Cai (Fujitsu), Liu Wenlong (Fujitsu), Yasuyuki Matsuura ==Topics== OLD: * Hardware test template for rpi3 https://git.automotivelinux.org/AGL/meta-agl/tree/templates/machine/raspberrypi3/test * can be replicated for porter & vayu (WIP: jsmoeller) * Stephane: weston startup to be worked on. Also looks into virtual screens for simulation purposes (e.g. including faking EDIDs). Tests must be independent of the compositor going forward. * Tests initially +/- or Fail/Pass * **Update**: move to weston 1.11, differences in startup of weston (weston-start) * poky/meta/recipes-graphics/wayland/weston-init/weston-start * wip for next meeting * Longer test reports need to be uploaded lateron to a frontend (which could be kernelci) * Tests should be wrapped e.g. with https://avocado-framework.github.io/ (e.g. TAP as output) * Can later be fed/reused in Fuego * **Update**: discussion moved to fuego, waiting for format decision * Target have tests scripts ready for next meeting * Build-up: * basic "service up" * then "client connect" * then "request/answer" NEW: * FUEGO BOF SLIDES: http://elinux.org/images/f/f6/Fuego-Status-and-Roadmap-ELC-2017-02.pdf * Work going on in '-next' branches: * Tim: https://bitbucket.org/tbird20d/ * Daniel: https://bitbucket.org/nirrognas/ <----- -next branch ! * Jan-Simon: https://bitbucket.org/dl9pf/ * We need to adapt AGL-JTA to the changes introduced by fuego's next branch ! * mainly changes in the job templates * support for board orchestrators * Test repo * https://gerrit.automotivelinux.org/gerrit/gitweb?p=src%2Fqa-testdefinitions.git;a=summary * small/short tests for CI cycle * Stephane: SDK/Docker container: * stable version for container / just weekly snapshot build right now * AI (dab): SDK for release * AI (jsmoeller): send stephane the release job * Matsuzawa: SDK - one SDK per board ? * Right now one SDK per board * reason is that every board uses slightly different optimization levels * SDK container can hold multiple SDK (and right now **per board**) * long-term solution: set DEFAULT_TUNE in our AGL "distro" conf file * e.g. https://github.com/Angstrom-distribution/meta-angstrom/blob/master/conf/distro/angstrom.conf#L112 * https://github.com/Angstrom-distribution/meta-angstrom/blob/master/conf/distro/include/arm-defaults.inc * needs to be applied likewise to arm32, arm64, x86-32, x86-64 * -> goal is to just have 4 SDKs * jsmoeller: question: does wgt format handle arch / noarch ? * Stephane: update container with jekyll dependencies (for doc) * Stephane: multiple SDK installation in container * jsmoeller: with the # of SDK reduced, we can ship the container right away * jsmoeller: we should also take a look at: * http://events.linuxfoundation.org/sites/events/files/slides/2017%20ELC%20Henry%20Bruce.pdf * http://events.linuxfoundation.org/sites/events/files/slides/2017%20ELC%20-%20Yocto%20Project%20Containers.pdf * http://events.linuxfoundation.org/sites/events/files/slides/2017%20ELC%20--%20Using%20devtool%20to%20Streamline%20your%20Yocto%20Project%20Workflow.pdf * AI towards DAB: * unification of the SDKs (aka DEFAULT_TUNE) (jsmoeller) * stable/release version of the container ==Next meeting== Mar 14th , same time ==== Meeting 2017/02/14 ==== ==Attendees== Jan-Simon, Stephane, Tsubone-san, Matsuzawa-san, Ronan, ==Topics== ==Note== * NEW: * AMM F2F : new jenkins templates deployed * Hardware test template for rpi3 https://git.automotivelinux.org/AGL/meta-agl/tree/templates/machine/raspberrypi3/test * can be replicated for porter () & vayu (jsmoeller) * Stephane: weston startup to be worked on. Also looks into virtual screens for simulation purposes (e.g. including faking EDIDs). Tests must be independent of the compositor going forward. * Tests initially +/- or Fail/Pass * Longer test reports need to be uploaded lateron to a frontend (which could be kernelci) * Tests should be wrapped e.g. with https://avocado-framework.github.io/ (e.g. TAP as output) * Can later be fed/reused in Fuego * Target have tests scripts ready for next meeting * Build-up: * basic "service up" * then "client connect" * then "request/answer" * OLD: * //Tests to run in CI builds//: * __short test for weston__ should also query resolution (as proof that not just service is started) * Stephane works on it https://jira.automotivelinux.org/browse/SPEC-305 * Update 2017/01/31: ?? * [[https://jira.automotivelinux.org/browse/SPEC-275]] * mid-term: run commands and __double-check all services__ are up * appfw, network (wifi/bluetooth, but depends on board), weston (see above) * Ronan looks at it. Resarching yocto's ptest. * Update 2017/01/31: ?? * __Benchmark__ - e.g. check stress (-c/-m/-i) * Jan-Simon * Update 2017/01/31: done (lava part) * Stephane: __KPI__ (key performance indicators) could be defined and applied e.g. to snapshots/release (reference is tizen mobile profile) * Push up SAT * SPEC-274, https://jira.automotivelinux.org/browse/SPEC-274 * Update 2017/01/31: ==Goals for next 2 weeks== ==Next meeting== ==== Meeting 2017/01/31 ==== ==Attendees== * Jan-Simon Moeller, Tim Bird, Matusura-san, Agustin ==Topics== * OLD: * //Tests to run in CI builds//: * __short test for weston__ should also query resolution (as proof that not just service is started) * Stephane works on it https://jira.automotivelinux.org/browse/SPEC-305 * Update 2017/01/31: ?? * [[https://jira.automotivelinux.org/browse/SPEC-275]] * mid-term: run commands and __double-check all services__ are up * appfw, network (wifi/bluetooth, but depends on board), weston (see above) * Ronan looks at it. Resarching yocto's ptest. * Update 2017/01/31: ?? * __Benchmark__ - e.g. check stress (-c/-m/-i) * Jan-Simon * Update 2017/01/31: done (lava part) * Stephane: __KPI__ (key performance indicators) could be defined and applied e.g. to snapshots/release (reference is tizen mobile profile) * Push up SAT * SPEC-274, https://jira.automotivelinux.org/browse/SPEC-274 * Update 2017/01/31: * Andrew: embedded-bits.co.uk: __sabre-lite__ can be used. V1 setup. * Update 2016/11/16: **Test Snapshot build done** * Update 2016/12/20: Job tempate WIP (jsmoeller) * Update 2017/01/19: Sabre broken on morty * Update 2017/01/31: still broken * __Code analysis__ - https://jira.automotivelinux.org/browse/SPEC-290 * Status: Assigned to Walt * Update 2017/01/31: WIP, Walt/Dan asking IT * https://jira.automotivelinux.org/browse/SPEC-354 * agl-image-weston FTBS * Assigned to Ronan * Update 2017/01/31: Image dependencies fixed. * New: * Lava-Fuego setup: https://jta.automotivelinux.org/fuego * Using -next branch from Daniel * Wrapper job brings-up Board over Lava before triggering batch job * Issues: we rely still on "local" ssh or vpn * Board is blocked quite long (build, deploy, exec) * right now either block the board until all is done * or each fuego test is an individual "lava job" * decouple the build from deploy/run * decouple trigger from batch, but flow control needed (e.g. batch job blocks) * AI (jsm) look into ftc on how to call single phases * TARGET_SETUP_LINK ... TARGET_TEARDOWN_LINK * Infrastructure update * New CI template done using MultiJob * https://build.automotivelinux.org/view/sandbox-meta-verify/ * CIB = **C**i **I**mage **B**uild * CIBT = **C**i **I**mage **B**oot **T**test ==Next meeting== Next F2F: * 2017-02-17 during AMM/F2F Next Call: * 2017-02-28 same time ==== Meeting 2017/01/19 ==== ==Attendees== * Jan-Simon Moeller, Walt Miner, Xiuhui Zhang, Yasuyuki Matsuura, ==Topics== * OLD: * //Tests to run in CI builds//: * __short test for weston__ should also query resolution (as proof that not just service is started) * Stephane works on it https://jira.automotivelinux.org/browse/SPEC-305 * Update 2016/12/06: **Update from Jens in Jira** * [[https://jira.automotivelinux.org/browse/SPEC-275]] * mid-term: run commands and __double-check all services__ are up * appfw, network (wifi/bluetooth, but depends on board), weston (see above) * Ronan looks at it. Resarching yocto's ptest. * Update 2016/12/06: **No update** * __Benchmark__ - e.g. check stress (-c/-m/-i) * Jan-Simon * Status update: work started, writing jjb templates * Update 2016/12/20: **No update, waiting for infra update** * Update 2017/01/19: jjb job done, can (build), upload, boot board, run test and upload result. * Stephane: __KPI__ (key performance indicators) could be defined and applied e.g. to snapshots/release (reference is tizen mobile profile) * Push up SAT * SPEC-274, https://jira.automotivelinux.org/browse/SPEC-274 * Update 2016/11/16: no kpi's available to reuse * __Docker container creation__ as part of jenkins builds e.g. monthly * JIRA to track - https://jira.automotivelinux.org/browse/SPEC-301 * AI(Jan-Simon): add to jenkins * Update: 2017/01/19: Done: https://download.automotivelinux.org/AGL/snapshots/sdk/docker/ * Andrew: embedded-bits.co.uk: __sabre-lite__ can be used. V1 setup. * Update 2016/11/16: **Test Snapshot build done** * Update 2016/12/20: Job tempate WIP (jsmoeller) * Update 2017/01/19: Sabre broken on morty * __Code analysis__ - https://jira.automotivelinux.org/browse/SPEC-290 * Status: Assigned to Walt * Update 2017/01/19: **WIP, no news** * https://jira.automotivelinux.org/browse/SPEC-354 * agl-image-weston FTBS * Assigned to Ronan * Update 2017/01/19: No news * Release log * repo diffmanifests default_blowfish_2.0.3.xml default_blowfish_2.0.4.xml * New: * Xiuhui Zhang working on JTA for porter/dragon * Matsuura-san looking into development environment and (test-)infra * Agustin gave update on CIP's progress on KernelCI frontend/backend VM. * Infrastructure update * YP Morty migration done. * Jenkins stable since start of jan. * But changed ulimits for jenkins slaves. ==Next meeting== 2017-01-31 same time ==== Meeting 2017/01/03 ==== ==Attendees== * Jan-Simon Moeller, Song Cai, ==Topics== * OLD: * //Tests to run in CI builds//: * __short test for weston__ should also query resolution (as proof that not just service is started) * Stephane works on it https://jira.automotivelinux.org/browse/SPEC-305 * Update 2016/12/06: **Update from Jens in Jira** * [[https://jira.automotivelinux.org/browse/SPEC-275]] * mid-term: run commands and __double-check all services__ are up * appfw, network (wifi/bluetooth, but depends on board), weston (see above) * Ronan looks at it. Resarching yocto's ptest. * Update 2016/12/06: **No update** * __Benchmark__ - e.g. check stress (-c/-m/-i) * Jan-Simon * Status update: work started, writing jjb templates * Update 2016/12/20: **No update, waiting for infra update** * Stephane: __KPI__ (key performance indicators) could be defined and applied e.g. to snapshots/release (reference is tizen mobile profile) * Push up SAT * SPEC-274, https://jira.automotivelinux.org/browse/SPEC-274 * Update 2016/11/16: no kpi's available to reuse * __Docker container creation__ as part of jenkins builds e.g. monthly * JIRA to track - https://jira.automotivelinux.org/browse/SPEC-301 * Stephane to send script to Jan-Simon to add to jenkins (16.04/slaves will need docker) * Update 2016/12/06: **Stephan sent script (github) ** * AI(Jan-Simon): add to jenkins * Update 2016/12/06: ** Working on it. Planned for Wed/Thu** * Andrew: embedded-bits.co.uk: __sabre-lite__ can be used. V1 setup. * Update 2016/11/16: **Test Snapshot build done** * Update 2016/12/20: Job tempate WIP (jsmoeller) * __Code analysis__ - https://jira.automotivelinux.org/browse/SPEC-290 * Status: Assigned to Walt * Update 2016/12/06: **WIP, no news** * https://jira.automotivelinux.org/browse/SPEC-354 * agl-image-weston FTBS * Assigned to Ronan * Release log * repo diffmanifests default_blowfish_2.0.3.xml default_blowfish_2.0.4.xml * New: * - * - * Infrastructure update ==Next meeting== ==== Meeting 2016/12/20 ==== ==Attendees== * Cancelled. ==Topics== * OLD: * //Tests to run in CI builds//: * __short test for weston__ should also query resolution (as proof that not just service is started) * Stephane works on it https://jira.automotivelinux.org/browse/SPEC-305 * Update 2016/12/06: **Update from Jens in Jira** * [[https://jira.automotivelinux.org/browse/SPEC-275]] * mid-term: run commands and __double-check all services__ are up * appfw, network (wifi/bluetooth, but depends on board), weston (see above) * Ronan looks at it. Resarching yocto's ptest. * Update 2016/12/06: **No update** * __Benchmark__ - e.g. check stress (-c/-m/-i) * Jan-Simon * Status update: work started, writing jjb templates * Update 2016/12/20: **No update, waiting for infra update** * Stephane: __KPI__ (key performance indicators) could be defined and applied e.g. to snapshots/release (reference is tizen mobile profile) * Push up SAT * SPEC-274, https://jira.automotivelinux.org/browse/SPEC-274 * Update 2016/11/16: no kpi's available to reuse * __Docker container creation__ as part of jenkins builds e.g. monthly * JIRA to track - https://jira.automotivelinux.org/browse/SPEC-301 * Stephane to send script to Jan-Simon to add to jenkins (16.04/slaves will need docker) * Update 2016/12/06: **Stephan sent script (github) ** * AI(Jan-Simon): add to jenkins * Update 2016/12/06: ** Working on it. Planned for Wed/Thu** * Andrew: embedded-bits.co.uk: __sabre-lite__ can be used. V1 setup. * Update 2016/11/16: **Test Snapshot build done** * Update 2016/12/20: Job tempate WIP (jsmoeller) * __Code analysis__ - https://jira.automotivelinux.org/browse/SPEC-290 * Status: Assigned to Walt * Update 2016/12/06: **WIP, no news** * https://jira.automotivelinux.org/browse/SPEC-354 * agl-image-weston FTBS * Assigned to Ronan * Release log * repo diffmanifests default_blowfish_2.0.3.xml default_blowfish_2.0.4.xml * New: * - * - * Infrastructure update ==Next meeting== ==== Meeting 2016/12/06 ==== ==Attendees== * Ronan, Stephan, Walt, Jan-Simon, Jian, toscalix ==Topics== * OLD: * //Tests to run in CI builds//: * __short test for weston__ should also query resolution (as proof that not just service is started) * combined with Blind tests e.g. report fps for display rendering (w/o actual display) * open, ask UI EG * SPEC-275 https://jira.automotivelinux.org/browse/SPEC-275 * weston should also start w/o monitor * Stephane works on it https://jira.automotivelinux.org/browse/SPEC-305 * investigating e.g. headless backend and framebuffer/fbcat * Stephane will try framebuffer/fbcat + compare first * what about wayland-fits ?? * Update 2016/12/06: **Update from Jens in Jira** * [[https://jira.automotivelinux.org/browse/SPEC-275]] * mid-term: run commands and __double-check all services__ are up * appfw, network (wifi/bluetooth, but depends on board), weston (see above) * Ronan looks at it. Resarching yocto's ptest. * first target is e.g. appfw * https://wiki.yoctoproject.org/wiki/Ptest * DISTRO_FEATURES_append = " ptest" * EXTRA_IMAGE_FEATURES += "ptest-pkgs" * run ptest-runner * All ptest files are installed in /usr/lib//ptest * a recipe adds a run-ptest * Add to -qa images * Update 2016/12/06: **No update** * __Benchmark__ - e.g. check stress (-c/-m/-i) * Jan-Simon * Status update: work started, writing jjb templates * lava yaml * lava_command_run: * Update 2016/12/06: **No update, waiting for infra update** * Stephane: __KPI__ (key performance indicators) could be defined and applied e.g. to snapshots/release (reference is tizen mobile profile) * Push up SAT * SPEC-274, https://jira.automotivelinux.org/browse/SPEC-274 * Stephan will see if there is doc from tizen avail * Update 2016/11/16: **No update** * __Docker container creation__ as part of jenkins builds e.g. monthly * JIRA to track - https://jira.automotivelinux.org/browse/SPEC-301 * Stephane to send script to Jan-Simon to add to jenkins (16.04/slaves will need docker) * Dockerfile/script * git repo for script (also JIRA) * Status: progress - changes for 2.2: * python3 * LANG setting (C.utf not supported, need to set e.g. en_US.utf-8) * Stephane will send updated script for CC * Standby * Update 2016/12/06: **Stephan sent script (github) ** * AI(Jan-Simon): add to jenkins * container might have to use debian/unstable (pseudo bug) for krogoth * tar, pbzip2, xzutils need newer versions * needs to be figured out * add to snapshots or weekly * repo in gerrit for agl-docker-worker (owner: sdx, ronan, jose) * make; make export * Andrew: embedded-bits.co.uk: __sabre-lite__ can be used. V1 setup. * JSON job needed * build for the sabre-lite * https://github.com/Freescale/meta-freescale-3rdparty/blob/master/conf/machine/imx6qsabrelite.conf * template * Update 2016/11/09: **DONE** * Token from LAVA instance * Update 2016/11/09: **DONE** * nbdroot patches (Jan-Simon to mail) * Status: **Sent** * Update 2016/11/16: **Test Snapshot build done** * Update 2016/12/06: Job tempate WIP * __Code analysis__ - https://jira.automotivelinux.org/browse/SPEC-290 * Status: Assigned to Walt * Update 2016/12/06: **WIP, no news** * New: * https://jira.automotivelinux.org/browse/SPEC-354 * agl-image-weston FTBS * can be used for simple UI tests * Assigned to Ronan * Release log * repo diffmanifests default_blowfish_2.0.3.xml default_blowfish_2.0.4.xml * Infrastructure update * JTA update to master WIP, redeploying/reconfiguring * Update 2016/11/16: **WIP, no news** * We need to wire-up the **Renesas Porter** and trigger test * 2nd step is to add the **TI vayu** * 3rd step is to add the **Intel Minnow** * **kvm**/qemux86-64 * LAVA infra being resetup for more boards beside porter * boards: * 1st set: porter, qemux86-64, raspberrypi * Status update: rpi setup, porter(update minion) and kvm(new setup) pending * Update 2016/12/06: **need to write code for NBD** * 2nd set: dra7xx-evm, dragonboard-410c, sabre ? * pipeline job migration * Status: **WIP** * satellite slave setup + documentation for setup of lab * Status: **No time, yet.** ==Next meeting== Dec 20th, same time, same place ==== Meeting 2016/11/16 ==== ==Attendees== ==Topics== * OLD: * //Tests to run in CI builds//: * __short test for weston__ should also query resolution (as proof that not just service is started) * combined with Blind tests e.g. report fps for display rendering (w/o actual display) * open, ask UI EG * SPEC-275 https://jira.automotivelinux.org/browse/SPEC-275 * weston should also start w/o monitor * Stephane works on it https://jira.automotivelinux.org/browse/SPEC-305 * investigating e.g. headless backend and framebuffer/fbcat * Stephane will try framebuffer/fbcat + compare first * what about wayland-fits ?? * Update 2016/11/16: **Update from Jens in Jira** * mid-term: run commands and __double-check all services__ are up * appfw, network (wifi/bluetooth, but depends on board), weston (see above) * Ronan looks at it. Resarching yocto's ptest. * first target is e.g. appfw * https://wiki.yoctoproject.org/wiki/Ptest * DISTRO_FEATURES_append = " ptest" * EXTRA_IMAGE_FEATURES += "ptest-pkgs" * run ptest-runner * All ptest files are installed in /usr/lib//ptest * a recipe adds a run-ptest * Add to -qa images * Update 2016/11/16: **No update** * __Benchmark__ - e.g. check stress (-c/-m/-i) * Jan-Simon * Status update: work started, writing jjb templates * lava yaml * lava_command_run: * Update 2016/11/16: **No update** * Stephane: __KPI__ (key performance indicators) could be defined and applied e.g. to snapshots/release (reference is tizen mobile profile) * Push up SAT * SPEC-274, https://jira.automotivelinux.org/browse/SPEC-274 * Stephan will see if there is doc from tizen avail * Update 2016/11/16: **No update** * __Docker container creation__ as part of jenkins builds e.g. monthly * JIRA to track - https://jira.automotivelinux.org/browse/SPEC-301 * Stephane to send script to Jan-Simon to add to jenkins (16.04/slaves will need docker) * Dockerfile/script * git repo for script (also JIRA) * Status: progress - changes for 2.2: * python3 * LANG setting (C.utf not supported, need to set e.g. en_US.utf-8) * Stephane will send updated script for CC * Standby * Update 2016/11/16: **No update** * Andrew: embedded-bits.co.uk: __sabre-lite__ can be used. V1 setup. * JSON job needed * build for the sabre-lite * https://github.com/Freescale/meta-freescale-3rdparty/blob/master/conf/machine/imx6qsabrelite.conf * template * Update 2016/11/09: **DONE** * Token from LAVA instance * Update 2016/11/09: **DONE** * nbdroot patches (Jan-Simon to mail) * Status: **Sent** * Update 2016/11/16: **Test Snapshot build done** * __Code analysis__ - https://jira.automotivelinux.org/browse/SPEC-290 * Status: Assigned to Walt * Update 2016/11/16: **WIP, no news** * NEW: * CIAT SPRINT: Thu Dec 1st 11:00 UTC or 14:00 UTC * * Infrastructure update * JTA update to master WIP, redeploying/reconfiguring * Update 2016/11/16: **WIP, no news** * We need to wire-up the **Renesas Porter** and trigger test * 2nd step is to add the **TI vayu** * 3rd step is to add the **Intel Minnow** * **kvm**/qemux86-64 * LAVA infra being resetup for more boards beside porter * boards: * 1st set: porter, qemux86-64, raspberrypi * Status update: rpi setup, porter(update minion) and kvm(new setup) pending * Update 2016/11/16: **WIP, instance up, boards wired-up, being connected now** * 2nd set: dra7xx-evm, dragonboard-410c, sabre ? * pipeline job migration * Status: **No time, yet.** * satellite slave setup + documentation for setup of lab * Status: **No time, yet.** ==Notes== ==Goals for next 2 weeks== ==Next meeting== ==== Meeting 2016/11/08 ==== ==Attendees== * Jan-Simon Möller * Walt Miner * Stephane Desneux * Yannic Gicquel ==Topics== * OLD: * Tests to run in CI builds: * short test for weston should also query resolution (as proof that not just service is started) * combined with Blind tests e.g. report fps for display rendering (w/o actual display) * open, ask UI EG * SPEC-275 https://jira.automotivelinux.org/browse/SPEC-275 * weston should also start w/o monitor * Stephane works on it https://jira.automotivelinux.org/browse/SPEC-305 * investigating e.g. headless backend and framebuffer/fbcat * Stephane will try framebuffer/fbcat + compare first * what about wayland-fits ?? * mid-term: run commands and double-check all services are up * appfw, network (wifi/bluetooth, but depends on board), weston (see above) * Ronan looks at it. Resarching yocto's ptest. * first target is e.g. appfw * https://wiki.yoctoproject.org/wiki/Ptest * DISTRO_FEATURES_append = " ptest" * EXTRA_IMAGE_FEATURES += "ptest-pkgs" * run ptest-runner * All ptest files are installed in /usr/lib//ptest * a recipe adds a run-ptest * Add to -qa images * Benchmark - e.g. check stress (-c/-m/-i) * Jan-Simon * Status update: work started, writing jjb templates * lava yaml * lava_command_run: * Stephane: KPI (key performance indicators) could be defined and applied e.g. to snapshots/release (reference is tizen mobile profile) * Push up SAT * SPEC-274, https://jira.automotivelinux.org/browse/SPEC-274 * Stephan will see if there is doc from tizen avail * Docker container creation as part of jenkins builds e.g. monthly * JIRA to track - https://jira.automotivelinux.org/browse/SPEC-301 * Stephane to send script to Jan-Simon to add to jenkins (16.04/slaves will need docker) * Dockerfile/script * git repo for script (also JIRA) * Status: progress - changes for 2.2: * python3 * LANG setting (C.utf not supported, need to set e.g. en_US.utf-8) * Stephane will send updated script for CC * Standby * Andrew: Intro + offers help . sabre-lite can be used. V1 setup. * JSON job needed * build for the sabre-lite * https://github.com/Freescale/meta-freescale-3rdparty/blob/master/conf/machine/imx6qsabrelite.conf * template * Not yet started * Token from LAVA instance * nbdroot patches (Jan-Simon to mail) * Status: Sent * Code analysis - https://jira.automotivelinux.org/browse/SPEC-290 * Status: Assigned to Walt- * NEW: * Schedule JTA/LAVA sprint - * Slot: this friday 4pm CET (UTC+1) ? * Otherwise travel due to F2F * Infrastructure update * JTA update to master WIP, redeploying/reconfiguring * LAVA infra being resetup for more boards beside porter * boards: * 1st set: porter, qemux86-64, raspberrypi * Status update: rpi setup, porter(update minion) and kvm(new setup) pending * 2nd set: dra7xx-evm, dragonboard-410c, sabre ? * pipeline job migration * Status: No time, yet. * satellite slave setup + documentation for setup of lab * Status: No time, yet. ==Next meeting== * We could do a call-in during the F2F next week. Pending schedule. I'll send an email out. * Likely thu 1700 JST / 0900 CET / 0800 UTC * 2016/11/22 same time ==== Meeting 2016/10/16 ==== ==Attendees== * Jan-Simon Möller * Walt Miner * Andrew Murray * Yuichi Kusakabe * Kengo Ibe * Yannic Gicquel * Stephane Desneux * Jian Zhang ==Topics== * OLD: * Tests to run in CI builds: * short test for weston should also query resolution (as proof that not just service is started) * weston should also start w/o monitor * e.g. like xrandr * e.g. framebuffer with specific resolution and run on framebuffer * Stephane * mid-term: run commands and double-check all services are up * appfw, network (wifi/bluetooth, but depends on board), weston (see above) * Stephane * Status update: no time, yet * Benchmark - e.g. check stress (-c/-m/-i) * Jan-Simon * Status update: not yet implemented * Stephane: KPI (key performance indicators) could be defined and applied e.g. to snapshots/release (reference is tizen mobile profile) * Push up SAT * SPEC-274, https://jira.automotivelinux.org/browse/SPEC-274 * Blind tests e.g. report fps for display rendering (w/o actual display) * open, ask UI EG * SPEC-275 https://jira.automotivelinux.org/browse/SPEC-275 * Docker container creation as part of jenkins builds e.g. monthly * JIRA to track - https://jira.automotivelinux.org/browse/SPEC-301 * Stephane to send script to Jan-Simon to add to jenkins (16.04/slaves will need docker) * Dockerfile/script * git repo for script (also JIRA) * Status: progress - changes for 2.2: * python3 * LANG setting (C.utf not supported, need to set e.g. en_US.utf-8) * Stephane will send updated script for CC * NEW: * Code analysis - https://jira.automotivelinux.org/browse/SPEC-290 * Andrew: Intro + offers help . sabre-lite can be used. V1 setup. * JSON job needed * build for the sabre-lite * https://github.com/Freescale/meta-freescale-3rdparty/blob/master/conf/machine/imx6qsabrelite.conf * template * e.g. for daily snapshots * JSON job file * Token from LAVA instance * nbdroot patches (Jan-Simon to mail) * JIRA: * Schedule JTA/LAVA sprint - Jan-Simon to send email/doodle . * Infrastructure update * Slaves see issues on sstate tar-native race, * "text file busy" * still investigating * bug submitted upstream - https://bugzilla.yoctoproject.org/show_bug.cgi?id=10449 * mitigation through automatic retriggers, but noisy in gerrit * Update downtimes on short notice * JTA update to master WIP, redeploying/reconfiguring * LAVA infra being resetup for more boards beside porter * boards: * 1st set: porter, qemux86-64, raspberrypi * 2nd set: dra7xx-evm, dragonboard-410c, sabre ? * pipeline job migration * satellite slave setup + documentation for setup of lab ==Next meeting== 2016/11/8 same time ==== Meeting 2016/10/04 ==== ==Attendees== * Jan-Simon Möller * Stephane Desneux * Mikhail Durnev * Yuichi Kusakabe * Kenji Tadano * Kengo Ibe * Koutaro Hashimoto * Yannic Gicquel * Walt Miner ==Topics== * Report/Summary of last weeks BOF session, discussion, tracking Jira's * Discussion: * short test for weston should also query resolution (as proof that not just service is started) * weston should also start w/o monitor * AI for today, who takes the lead on this ? JIRA entry, assign to. * open * mid-term: run commands and double-check all services are up * AI for today, who takes the lead on this ? JIRA entry, assign to. * appfw, network, weston * Stephane * Benchmark - e.g. check stress (-c/-m/-i) * AI for today, who takes the lead on this ? JIRA entry, assign to. * Jan-Simon * Stephane: KPI (key performance indicators) could be defined and applied e.g. to snapshots/release (reference is tizen mobile profile) * AI for today, who takes the lead on this ? JIRA entry, assign to. * Push up SAT * Blind tests e.g. report fps for display rendering (w/o actual display) * AI for today, who takes the lead on this ? JIRA entry, assign to. * open, ask UI EG * Nuohan to ask Agustin if there are tests/testcases executed by genivi/gdp * News ? * "As I mentioned in the BoF as soon as we finish the first set and we test they are working in production, we will publish them." * New: * Docker container creation as part of jenkins builds e.g. monthly * JIRA to track * Stephane to send script to Jan-Simon to add to jenkins (16.04/slaves will need docker) * Dockerfile/script * git repo for script (also JIRA) * Infrastructure update * Infra back up * CI builds in place * little bumpy during migration - sry! * expect smaller hickups * snapshot builds in place * test run triggered through snapshot builds * blowfish release build ==Goals for next weeks== * Work on AI * Docker container * ? JTA release for blowfish ==Next meeting== 2016/10/11 same time ==== Meeting 2016/09/13 ==== ==Attendees== * Jan-Simon Möller * Nuohan Qiao * Stephane Desneux * Kenji Tadano * Kengo Ibe * Koutaro Hashimoto ==Topics== * Roll call * Status round and AIs: * Report/Summary of last weeks BOF session, discussion, tracking Jira's * Discussion: * short test for weston should also query resolution (as proof that not just service is started) * weston should also start w/o monitor * mid-term: run commands and double-check all services are up * Benchmark - e.g. check stress (-c/-m/-i) * Stephane: KPI (key performance indicators) could be defined and applied e.g. to snapshots/release (reference is tizen mobile profile) * Blind tests e.g. report fps for display rendering (w/o actual display) * Nuohan to ask Agustin if there are tests/testcases executed by genivi/gdp * Infrastructure update * gerrit migration this week * jenkins setup finalized, to be migrated soon ==Next meeting== * 27th Sept, usual time ==== Meeting 2016/08/30 ==== ==Attendees== * Jan-Simon Möller * Nuohan Qiao * Stephane Desneux * Kenji Tadano * Kengo Ibe * Koutaro Hashimoto * Mikhail Durnev * Yuichi Kusakabe * Walt Miner ==Topics== * Roll call * collecting topic additions * -tbd- * Status round and AIs: * Other topics, open round: * --tbd-- * meta-isafw use and reporting, how to deal with error reports, submit bugs * Log bugs * Focus on static analysis * File bug reports * Include additions for ARM arch (static analysis) * Submit report in jira / plugin (jenkins/jira)? * Mail out report * AI: do meta-isafw run for BB rel 2.0.2 * AI: setup ML to send out reports * Plan for (post) release of JTA matching 'B' (shortly after 'B' release) - status update: * optional lava integration for boards (Jan-Simon/Nuohan) * definition of 3 test categories (short/mid/long) (Nuohan) * Done. Todo: define what tests should run in these categories -> BOF@AMM * config templates for porter and minnow (Nuohan) * porter in git * minnowboard to be added * dragonboard (close to done) * Support for multiple boards of same type (use lava for farm, request free board from there) * installation documentation for 2 use-cases (AGL-JTA solo, AGL-JTA & lava) (Jan-Simon, pending) * wiki: https://wiki.automotivelinux.org/agl-testframework * Infrastructure update: * LDAP migration DONE * This week gerrit * Next 2 weeks jenkins/download * AI: update WIKI pages wrt login on identity.lf.org ==Goals for next 2 weeks== * Infra work * Sprint 12th Sept 8am UTC ==Next meeting== * 13th Sept, usual time. ==== Meeting 2016/08/16 ==== ==Attendees== * Jan-Simon Möller * Nuohan Qiao ==Topics== * Roll call * collecting topic additions * -tbd- * Status round and AIs: * Other topics, open round: * --tbd-- * meta-isafw use and reporting * Plan for (post) release of JTA matching 'B' (shortly after 'B' release) - status update: * optional lava integration for boards (Jan-Simon/Nuohan) * definition of 3 test categories (short/mid/long) (Nuohan) * Done. TODO: actually assign which tests to run * config templates for porter and minnow (Nuohan) * porter in git * minnowboard to be added * dragonboard (?) * installation documentation for 2 use-cases (AGL-JTA solo, AGL-JTA & lava) (Jan-Simon) (pending) * wiki: https://wiki.automotivelinux.org/agl-testframework * Infrastructure update: * This week gerrit * Next 1-2 weeks, jenkins+download ==Goals for next 2 weeks== * BOF session @AMM (Nuohan will lead/moderate - thanks!) ==Next meeting== * 13th September, usual time. ==== Meeting 2016/08/02 ==== ==Attendees== * Nuohan Qiao * Jan-Simon Möller * Kengo Ibe * Kotaro Hashimoto * Walt Miner * Yannic Gicquel ==Topics/Notes== * Roll call * collecting topic additions * -tbd- * Status round and AIs: * Plan for (post) release of JTA matching 'B' (shortly after 'B' release) * optional lava integration for boards (Jan-Simon/Nuohan) * definition of 3 test categories (short/mid/long) (Nuohan) * We to select what test make sense for short/mid/long (10min/1h/6h) * config templates for porter and minnow (Nuohan) * porter in git * minnowboard to be added * dragonboard (?) * installation documentation for 2 use-cases (AGL-JTA solo, AGL-JTA & lava) (Jan-Simon) (pending) * wiki: https://wiki.automotivelinux.org/agl-testframework * AI 2 from last week: Test integration - pending * jta.automotivelinux.org * ( porter.automotivelinux.org (lava) ) * TBD switch on for 'BB' release builds * Proposal sprint calls/gotomeetings on: * 4th Aug (time to be determined by doodle) * http://doodle.com/poll/ah3pb3zbfa59ycmm * 12th Aug (time to be determined by doodle) * http://doodle.com/poll/y56kzfgha8k3mvuv * Other topics: * --tbd-- * Infrastructure update: * LDAP migration next * then gerrit/jenkins/download * Diplay of test results * Easy way for failure detection * For project, trend detection * in gerrit: CIBT CIB CILT CIUT ==Goals for next 2 weeks== * release of JTA matching 'B' (Nuohan, Jan-Simon) * config templates for porter and minnow (Nuohan) * porter (go through test and open AI's) (Yannic/Stephane) * minnowboard () * dragonboard () ==Next meeting== * 16th August, usual time. ==== Meeting 2016/07/19 ==== ==Attendees== * Jan-Simon Möller * Yannic Gicquel * Kenji Tadano * Kengo Ibe * Kotaro Hashimoto * Mikhail Durnev * Agustin Benito Bethencourt * Walt Miner ==Topics== * Roll call * collecting topic additions * * Status round and AIs: * ALS feedback * Todo for 'B' release * Integration LAVA+JTA for porter board (Jan-Simon & Nuohan) * Release branches, CI builds for release branches * Branches done * Builds WIP, blocked by fetcher issues * Plan for release of JTA matching 'B' (shortly after 'B' release) * optional lava integration for boards (Jan-Simon) * definition of 3 test categories (short/mid/long) (Nuohan) * config templates for porter and minnow (Nuohan) * installation documentation for 2 use-cases (AGL-JTA solo, AGL-JTA & lava) (Nuohan, Jan-Simon) * wiki: https://wiki.automotivelinux.org/agl-testframework * "Test roles and responsibilities", deferred after BB, discuss on ML and SAT * AIs from last meeting: * AI 1: Templates - done * AI 2: Test integration * jta.automotivelinux.org * ( porter.automotivelinux.org (lava) ) * TBD switch on for 'BB' release builds * Infrastructure update: * LDAP migration next * then gerrit/jenkins/download ==Notes== * ALS feedback: * Fuego Session, challenges * Fujitsu showed what had been done * Need for collaboration * UI modifications needed to update jenkins * Short-term focus in collab/merging codebase with fuego ==Goals for next 2 weeks== * Integrate JTA in release builds (full run, report only, no fail) ==Next meeting== * 2nd August, usual time. ==== Meeting 2016/07/05 ==== ==Attendees== * Jan-Simon Möller * Nuohan Qiao * Yannic Gicquel * Kenji Tadano * Kengo Ibe * Koutaro Hashimoto * Stephane Desneux * Walt Miner * Mikhail Durnev * Yury Asheshov ==Topics== * Roll call * collecting topic additions * -- * Status round and AIs of last two week: * Template system rework (Jan-Simon, Stephane) * Collection AIs to be done before 'C' release * JTA container for porter (Jan-Simon) * Integration LAVA+JTA for porter board (Jan-Simon & Nuohan) * Release branches, CI builds for release branches (TBD) * Images for download next week, tbd ==Notes== * collecting topic additions * -- * Status round and AIs of last two week: * Template system rework (Jan-Simon, Stephane) * Collection AIs to be done before 'B' release * JTA container for porter (Jan-Simon) * https://jta.agl.homelinux.org/ * Integration LAVA+JTA for porter board (Jan-Simon & Nuohan) * Release branches, CI builds for release branches (TBD) * Images for download next week, tbd ==Goals for next 2 weeks== * AI 1: Templates * AI 2: Test integration ==Next meeting== * F2F during ALS * 19th July, usual time. ==== Meeting 2016/06/22 ==== ==Attendees== * Jan-Simon Möller * Nuohan Qiao * Yannic Gicquel * Kengo Ibe * Mikhail Durnev * Koutaro Hashimoto * Stephane Desneux * Walt Miner ==Topics== * Roll call * collecting topic additions * -- * Status round and AIs of last two week: * JTA update * AI: Function.* recipes * Collection AIs to be done before 'C' release * JTA container for porter * export container as archive * Integration LAVA+JTA for porter board * Release branches, CI builds for release branches * rc snapshots? * weekly snapshot/daily snapshot with qa-test * promotion of weekly to milestone when new feature enters tree * Feature in staging, evaluation with builds, we can't demo features easily * bmaptool index creation * Tar usually too old for security labels (tar >= 1.28 required) ==Notes== * Roll call * collecting topic additions * * Status round and AIs of last two week: * JTA update * Categorizing is done * New features: - list of testresults * Repo in staging to host test results * AI: Function.* recipes * OpenSSL, whetstone + rmaptest left (Jan-Simon) * Collection AIs to be done before 'C' release * JTA container for porter * export container as archive * Integration LAVA+JTA for porter board * Release branches, CI builds for release branches * rc snapshots? * weekly snapshot/daily snapshot with qa-test * promotion of weekly to milestone when new feature enters tree * Feature in staging, evaluation with builds, we can't demo features easily * experimental/devel branch ? * build based on topic in gerrit ? * push to sandbox everywhere and adapt manifest in sandbox of AGL-repo * bmaptool index creation * class in yocto ? * recipe / native for bmaptool * Tar usually too old for security labels (tar >= 1.28 required) * tar-native ? * publish sdcard images ! (porter) ==Next meeting== Tuesday July 5th - usual time. ==== Meeting 2016/06/30 ==== ==Attendees== * Nuohan Qiao * Kengo Ibe * Mikhail Durnev * Kenji Tadano * Koutaro Hashimoto * Stephane Desneux * Walt Miner * Jan-Simon Möller * Paul W. Nichols * Shotaro Arakawa * Agustin Benito Bethencourt ==Topics== * Roll call * collecting topic additions * -- * Status round and AIs of last two week: * Integrate nbdroot to meta-agl/meta-netboot * Security framework/LSM and nfsroot (Stephane) [[https://jira.automotivelinux.org/browse/SPEC-175|SPEC-175]] * AI: Investigate how to build meta-renesas w/o blob (limited func) (Stephan) * JTA update * meta-isafw * Integrated now, but how to handle the logs * db or upload or what ?! * AI: Function.* recipes (Stephane, Jan-Simon, Yannic) * AGL Jenkins update * Open round * End ==Notes== * Roll call * collecting topic additions * None * Status round and AIs of last two week: * Integrate nbdroot to meta-agl/meta-netboot * modified revision required INHERIT += "netboot" * fixed issue with minnowboard max * u-boot based boards should work now * uefi-based setup needs checking (minnox max) * tftp client on minnowboard max is broken * e.g. use ipxe or chainload into u-boot * Integration for nbdroot in lava * builds / test-boot with smack enabled * https://porter.agl.homelinux.org/scheduler/job/75 * Target for security framework integration is before BB release for demo (mid july) * documentation to me released next week * enable basic features of application framework (qemu + porter + (vayu) ) * JTA update * Categorization of testcases is wip * New Feature wip: list test results * meta-isafw * manual process, inspection * start from snapshots, then move-on to release branches/rc's only * AI: Function.* recipes (Stephane, Jan-Simon, Yannic) * ETA - 3 weeks * AI: Jan-Simon send remaining list to AGL ML * AGL Jenkins update * Open round * End ==Goals for next 3 weeks== * Function.* recipes * Enable security framework in builds (qemu+porter[+vayu]) before ALS ==Next meeting== 2016-06-22 (=Wed) @ the usual time (12:00 UTC) ==== Meeting 2016/05/17 ==== ==Attendees== * Nuohan Qiao * Kengo Ibe * Mikhail Durnev * Kenji Tadano * Koutaro Hashimoto * Agustin Benito Bethencourt * Yuichi Kusakabe * Jan-Simon Möller ==Topics== * Roll call * collecting topic additions * None * Status round and AIs of last two week: * Integrate nbdroot to meta-agl/meta-netboot * Security framework/LSM and nfsroot (Stephane) [[https://jira.automotivelinux.org/browse/SPEC-175|SPEC-175]] * AI: Investigate how to build meta-renesas w/o blob (limited func) (Stephan) * JTA update * JTA results in git repo * meta-isafw * Integrated now, but how to handle the logs * db or upload or what ?! * AI: Function.* recipes (Stephane, Jan-Simon, Yannic) * AGL Jenkins update * Open round * End ==Minutes== * Status round and AIs of last two week: * Integrate nbdroot to meta-agl/meta-netboot * Security framework/LSM and nfsroot (Stephane) [[https://jira.automotivelinux.org/browse/SPEC-175|SPEC-175]] * AI: Investigate how to build meta-renesas w/o blob (limited func) (Stephan) * JTA update * Question: Public JTA, currently deferred due to LF IT busy with gerrit and main jenkins install. * simple solution can be git, but no mid/long-term * staging/test-results.git with subfolder for AGL-JTA and e.g. (meta-)isafw (and other tests) * Need to check if debian as host is possible (but usual OS is CentOS) * benchmark.netperf failing * [[https://git.automotivelinux.org/gerrit/#/c/5653/| c5653]] - categories for JTA * Feedback needed if change works for you * meta-isafw * Integrated now, but how to handle the logs * db or upload or what ?! * Same issue as with AGL-JTA test output. * e.g. database + webui (django e.g. kernelci) * AI: Function.* recipes (Stephane, Jan-Simon, Yannic) * More tests included not, few left * Soon try to skip build in AGL-JTA for the included tests. * AGL Jenkins update * Installation done, next is the Jenkins Job Builder * IT is providing a new gerrit (data will be migrated) * LTP update in AGL-JTA, also has recipe in oe-core, we use version LTP 20160510 * recipe or bbappend for us (backport to jethro) * Open round * No more topics * End * Meeting ended 12:45 UTC ==Goals for next 2 weeks== * AI: Function.* recipes (Stephane, Jan-Simon, Yannic) * AI: jansimon: Need to check if debian as host is possible (but usual OS is CentOS) * AI: LTP recipe update * recipe or bbappend for us (backport to jethro) ==Next meeting== Next meeting in 2 weeks 2016-05-31 (=Tue) @ the usual time (12:00 UTC) ==== Meeting 2016/05/03 ==== ==Attendees== * Yannic Gicquel * Stephane Desneux * Paul W. Nichols * Jan-Simon Möller * Walt Miner * Kenji Tadano * Kotaro Hashimoto ==Topics== * Roll call * collecting topic additions * Status round and AIs of last two week: * Security framework/LSM and nfsroot (Stephane) [[https://jira.automotivelinux.org/browse/SPEC-175|SPEC-175]] * AI: Investigate how to build meta-renesas w/o blob (limited func) (Stephan) * AI: Function.* recipes (Stephane, Jan-Simon, Yannic) * JTA update * Jenkins update * Discussion on repo vs. combo-layer tooling and +/- (also topic in dev call) * Open round * End ==Minutes== * Roll call * collecting topic additions * no additions * Review of AIs * Stephane managed to boot and AGL image with Initrd into a nbd rootfs * Boot works from initrd (full filesystem, not initramfs) * initscript runs (busybox image) * initscript mounts base filesystems, detects smack and does further setup * userspace client needs to be started with appropriate rights * switch to new root and executing systemd * Developer setup: export deploy directory with tftp for kernel/initrd -> huge time win * nbd was simplest choice in our eval - easy to integrate * currently hosted in meta-netboot * filesystem image from yocto (ext3/ext4) can be used * uInitrd and Initrd should be supported by the layer * Review of the test AIs * JTA update: currently working on native deployment * Jenkins: IT creates cloud deployment for Jenkins right now. Migration asap. * Discussion on repo vs. combo-layer tooling and +/- (also topic in dev call) * Deferred to dev call * Open round * meta-security-isafw - Stephane to submit layer to AGL-repo. Test with porter snapshot. * Collecting next AIs * End * The meeting ended 12:45 UTC ==Goals for next 2 weeks== * Integrate nbdroot to meta-agl/meta-netboot * Functional.* * Left items: * Stephane: * Functional.aiostress - wip, extra sources * Functional.arch_timer (=dung) - wip, extra sources * Functional.fontconfig - recipe exists, need to double-check + package-group * Functional.fsfuzz - recipe exists, need to double-check + package-group * Functional.ft2demos - recipe exists, need to double-check + package-group * Yannic: * Functional.glib - ptest / distro feature avalable * Functional.ipv6connect - chromium related, source to be copied * Functional.jpeg - already packaged * Functional.linus_stress - chromium related, source to be copied * Functional.mesa-demos - x11, so not applicable * Functional.pi_tests - part of rt-test * Jan-Simon: * Benchmark.lmbench2 - lmbench3 (but build issue) * Benchmark.OpenSSL - recheck with openssl in yocto's ptest * Benchmark.Whetstone - write sources + recipe * Functional.crashme - ?? scrashme/trinity * Functional.rmaptest - license ?! * meta-security-isafw -> porter snapshot ==Next meeting== Next meeting in 2 weeks 2016-05-17 (=Tue) @ the usual time (12:00 UTC)- ----- ==== Meeting 2016/04/18 ==== ==Attendees== * Nuohan Qiao * Yannic Gicquel * Stephane Desneux * Kengo Ibe * Kenji Tadano * Rimpei Sosa * Jan-Simon Möller * Mikhail Durnev * Walt Miner * Shotaro Arakawa * Yuichi Kusakabe ==Topics== * Roll call * collecting topic additions * Status round and AIs of last week: * JTA fork, jenkins update (Nuohan) * AI: set up new JTA from new codebase * AI: Investigate how to build meta-renesas w/o blob (limited func) (Stephan) * AI: Function.* recipes (Stephane, Jan-Simon, Yannic) * AI: patch review * AI: investigate Security/LSM/nfsroot * Security framework/LSM and nfsroot (Stephane) ==Minutes== * Topic additions * Status round * JTA: * Almost done, uploading patches this week * Jan-Simon to start new container once patches landed (end of next week) * Manual for JTA on real machine (instead of docker) in the queue * AI's: * Snapshot / populate_sdk: * https://download.automotivelinux.org/AGL/snapshots/master/2016-04-18-b211/qemux86-64/deploy/sdk/ * AI: Stephane: use SDK in docker * AI: Investigate how to build meta-renesas w/o blob (limited func) (Stephan, Yannic) * AI: Function.* recipes ... requeued * Yannic: repo for test sources: * src/qa-test-misc created * Security/LSM/nfsroot * Issue appeared when connman was introduced, solved for now * BUT with active LSM nfsroot will break again (NFS has no ext. attr) * Timeframe ~ June for activation of security framework * Ideas: * use initrd + iSCSI/nbd/AoE/xyz to work w/o * likely changes in the test frameworks * board with usb-OTG port could server as programmable usb-key * Linaro uses sdcard-boards attached to 2 machines (but slow!) * see/discuss/comment on Jira SPEC-175 * Discussion on repo vs. combo-layer tooling and +/- ==Goals for next 2 weeks== * AI: Function.* recipes (Stephane, Jan-Simon, Yannic) * AI: patch review * AI: investigate Security/LSM/nfsroot * AI: set up new JTA from new codebase ==Next meeting== Next meeting in 2 weeks 2016-05-03 (=Tue) @ the usual time ==== Meeting 2016/04/5 ==== ==Attendees== * Walt Miner * Nuohan Qiao * Yannic Gicquel * Stephane Desneux * Munakata-san * Agustin Benito Bethencourt ==Topics== * Roll call * collecting topic additions * open status round * Status of AIs of last week: * AI: recipes for JTA-core components * Stephane, Jan-Simon, Nuohan, Yannic * Stephan & Yannic finished the packages, Jan-Simon has 3 left, rest is packagegroup * AI: Snapshot should write out the populate_sdk (jsmoeller) * AI: Nuohan JTA * AI: Investigate how to build meta-renesas w/o blob (limited func) * what about core-image-minimal(-qa) (Stephane) * current: there is still a dependency pulling in the blobs (agl-image-minimal) * JTA fork, jenkins update * JTA-core proposal (Nuohan) * Nouhan: Updated jenkins and ran bzip test. Estimate is about 2 week. Will send update to ML. * Munakata-san: proposes collaboration with LTSI to pull them along the codebase. * we should also work together with fuego * Image build porter board options * Either core-image-minimal * Or remove or mask the blobs (? is there a bitbake option) * Plans for the next 2 weeks, ownership of topics * AI: Snapshot should write out the populate_sdk (jsmoeller) * AI: Investigate how to build meta-renesas w/o blob (limited func) (Stephan) * AI: Function.* recipes (Stephane, Jan-Simon, Yannic) ==Notes== ==Next meeting== Next meeting in 2 weeks @ the usual time ==== Meeting 2016/03/22 ==== ==Attendees== * Walt Miner * Nuohan Qiao * Agustin Benito Bethencourt * Paul Sherwood * Shotaro Arakawa * Stephane Desneux * Yuichi Kusakabe * Sosa Rimpei * Paul Nichols * Yannic Gicquel * Jan-Simon Möller ==Topics== * Roll call * CIAT EG lead election * Status update * New gerrit repos, ACLs * Nuohan is updating the jenkins version of JTA, UI simplification to all for easier upgrades * Topics from last meeting: * Contributed repos, on-demand CI builds * /src and /staging available in gerrit * request repo creation through jira ticket * JTA fork, jenkins update * Tests compiled as part of AGL distro * AI: Next step is to find/import/write recipes in meta-agl to build the test. * Image build porter board options to be discussed: * splitting binary part apart, ship w/o, user can post-install. * stub out with dummy libraries * build w/o proprietary parts * Single image build (e.g. also qemuarm for meta-renesas) * Update by Stephane: * wrapper for SDK wip to build layer with * Plans for the next 2 weeks, ownership of topics * AI: recipes for JTA-core components * Stephane, Jan-Simon, Nuohan, Yannic * AI: Snapshot should write out the populate_sdk (jsmoeller) * AI: Nuohan JTA * AI: Investigate how to build meta-renesas w/o blob (limited func) * what about core-image-minimal(-qa) (Stephane) == Next meeting == * Due to the ELC, we might consider doing the call a little later and bridge F2F and the gotomeeting bridge. 9am pacific might work for those attending ELC (8am pacific might also do, but not sure about the room - will investigate). * We keep tuesday in 2 weeks and plan for a combined F2F + gotomeeting, invite by email the day before. ---- ==== Meeting 2016/03/09 ==== ==Attendees== * Walt, Nuohan, Stephane, Yannic, Bai, Jan-Simon ==Topics== * Status round * JTA discussion, maintainance of JTA unclear, should AGL take over * Implementation tbd, e.g. in container environment triggered by main jenkins * Test scheduler tbd * JTA, LAVA(board management)+JTA, LAVA * Test should be compiled as part of AGL distro * Test case tool/scheduler intro (Stephane) * Test case definition for use-cases * e.g. gerrit changeset (short/sanity) * snapshot (medium) * ( e.g. weekly) * release (long) * AGL Spec 2.0 Status/Review (Walt) * (check connections wrt software update, package signing) * Meeting schedule discussion * bi-weekly * time 12 UTC * Tuesday * CIAT EG team lead - proposals and election process * Proposals by email (to Walt or mailinglist) * Contributed repos, on-demand CI builds * Image build porter board * options to be discussed: * splitting binary part apart, ship w/o, user can post-install. * stub out with dummy libraries * build w/o proprietary parts * Single image build (e.g. also qemuarm for meta-renesas) * Idea: f2f with ppl attending ELC ==Next meeting== * The next meeting is scheduled for: 22nd mar 2016 12:00 UTC ==== AGL AMM 2016 ==== EG session during the AGL AMM 2016 (winter): The present members of the CIAT EG met during AGL AMM 2016. For the next cycle these 4 goals of the EG were identified: === Goals === * Overall the board support and coverage of the tests are to be increased. Therefore we need primarily open source boards as test targets as only those image builds can be shared/distributed. This is an essential requirement for running collaborative and effective public testing for AGL. * 3-4 working topics for the next iteration towards Brilliant Blowfish * Upgrade of 'embedded' Jenkins in JTA and some plugins * The current version of Jenkins within JTA is very old (2013). To host a public instance, we need a current Jenkins. * Split of JTA-Core * JTA-Core is the set of scripts that run the tests. Splitting them from JTA allows us to reuse it. * Run JTA-Core within Lava * With the tests separated from JTA, we can run it in other frameworks like lava. * Is it possible to include the tests (e.g. jta-core) in the AGL distro build as recipes (See/investigate LTP test recipe from ...) * JTA assumes that the tests need to be compiled before they can run. We can speed up the test procedure if the tests are already compiled within the AGL distro as packages.