11438 2017-05-02 10:43:00 +0000 oeqa: QemuRunner improvements 2020-10-22 07:09:18 +0000 1 1 10 QA/Testing Build Testing general unspecified x86 Multiple RESOLVED OBSOLETE Medium enhancement Future 1 mikko.ylinen unassigned benjamin.esquivel patrick.ohly randy.macleod richard.purdie --- --- No (bug/feature does not impact docs) oldest_to_newest 72978 0 mikko.ylinen 2017-05-02 10:43:00 +0000 oeqa's QemuRunner needs a few enhancements/cleanups, for example, to make it more suitable for ovmf (EFI Shell) testing with Qemu. Background: since ovmf was added in OE-core, it opens us up more opportunities to do, e.g., UEFI secure boot testing using Qemu. TODO list: - It must be possible to run commands in EFI shell too (see ovmf-shell-image). Currently, QemuRunner fails if it does not find regular login prompt (http://git.openembedded.org/openembedded-core/tree/meta/lib/oeqa/utils/qemurunner.py#n277) and if the login fails (http://git.openembedded.org/openembedded-core/tree/meta/lib/oeqa/utils/qemurunner.py#n301). This logic must be made more flexible so that qemu connection is also available in EFI shell ('Shell>'). Perhaps implement 'autologin' imagefeature (See Refkit) and start using that. - These prints are redundant http://git.openembedded.org/openembedded-core/tree/meta/lib/oeqa/utils/qemurunner.py#n291 because qemulog is printed out later. - targetcontrol.py shall not depend on bb. Things like http://git.openembedded.org/openembedded-core/tree/meta/lib/oeqa/targetcontrol.py#n201 are better in logger. 88453 1 richard.purdie 2020-10-22 07:09:18 +0000 I checked and I think some of the latter generic pieces were sorted out. The EFI issues probably remain but would need someone coming at it from a fresh perspective now, this bug isn't ageing well. Marking as obsolete as someone would come at this from a fresh perspective now.