| Summary: | meta-intel do_testimage tasks failed | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Build System, Metadata & Runtime] BSPs | Reporter: | Amanda Brindle <amanda.r.brindle> | ||||
| Component: | bsps-meta-intel | Assignee: | Yogesh Tyagi <yogesh.tyagi> | ||||
| Status: | IN PROGRESS IMPLEMENTATION --- | QA Contact: | |||||
| Severity: | normal | ||||||
| Priority: | Medium | CC: | chee.yang.lee, randy.macleod, richard.purdie, ross.burton, tim.orling | ||||
| Version: | unspecified | ||||||
| Target Milestone: | 6.99 | ||||||
| Hardware: | x86 | ||||||
| OS: | Multiple | ||||||
| Whiteboard: | |||||||
| OS type for building Yocto: | --- | Type of Regression: | --- | ||||
| Verified: | Documentation change: | No (bug/feature does not impact docs) | |||||
| Attachments: |
|
||||||
|
Description
Amanda Brindle
2019-09-24 18:09:57 UTC
Anuj dug and we concluded that the problem is that the intel machines don't use the hardcoded networking configuration that we expect for testimage. I've a partial fix for this locally. Downgrading to M+ because this isn't a release blocker. The problem is that currently testimage assumes that the machines being booted are the qemu machines (qemuarm, qemux86, etc) and these have specific a network configuration set through the connman-conf recipe with its qemuall overrides. As meta-intel BSPs don't have this the configuration is different and testimage can't find the machine once it has booted. I can see several potential solutions and likely more: 1) extract the qemu-specific bits of connman-conf to a new recipe connman-conf-testimage, and install that in all images which will be ran in testimage (maybe testimage.bbclass could append IMAGE_INSTALL). 2) run a DHCP server on the virtual network device and hand out a static IP instead of hardcoding an IP A start at (1) is in poky-contrib:ross/testnet but it isn't complete in any way. Reassigning to Anuj as the meta-intel owner. Bulk move from 4.99 or 0.00 to 5.99 Yogesh, Is this still a valid issue for meta-intel ? |