| Summary: | runqemu scripts should attempt sane defaults for unknown machines | ||
|---|---|---|---|
| Product: | [Build System, Metadata & Runtime] Meta-yocto | Reporter: | Darren Hart <dvhart> |
| Component: | meta-yocto | Assignee: | Unassigned <unassigned> |
| Status: | RESOLVED WONTFIX | QA Contact: | |
| Severity: | enhancement | ||
| Priority: | Low | CC: | bluelightning, jessica.zhang, jon.mason, poky.bs.watcher, poky.watcher, randy.macleod, scott.a.garman, sgw, song.liu |
| Version: | unspecified | ||
| Target Milestone: | Future | ||
| Hardware: | All | ||
| OS: | Multiple | ||
| Whiteboard: | |||
| OS type for building Yocto: | --- | Type of Regression: | --- |
| Verified: | Documentation change: | Yes (doc changes required) | |
This should in theory be easier with the newly rewritten runqemu that uses the QB_* variables - we could presumably set defaults for these on a per-architecture basis. We could guess about this required options but we could easily be wrong. Better to be deterministic. |
When building for certain real machines, it can still useful to test the results in qemu. Currently, runqemu and runqemu-internal must have several things specified for every machine, and if not, abort with an error. Instead, it would be nice if it would at least attempt to run with sane defaults: FS=ext3 KERNEL={bzImage,uImage,etc.}-$MACHINE.bin The qemu${ARCH} might be guessable from the KERNEL image, and if not, an ARCH= option might be a reasonable addition. As for the QEMUOPTIONS, perhaps a serial port for output rather than VGA if that is easier to deal with.