Bug 4090 - MACHINEOVERRIDES still inconsistent
Summary: MACHINEOVERRIDES still inconsistent
Status: RESOLVED FIXED
Alias: None
Product: OE-Core
Classification: Build System, Metadata & Runtime
Component: configuration (show other bugs)
Version: 1.3
Hardware: All Multiple
: Undecided normal
Target Milestone: ---
Assignee: Richard Purdie
QA Contact:
URL:
Whiteboard:
Depends on:
Blocks:
 
Reported: 2013-03-24 12:38 UTC by Jochen Trumpf
Modified: 2013-03-25 20:52 UTC (History)
0 users

See Also:
OS type for building Yocto: ---
Type of Regression: ---
Verified:
Documentation change: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Jochen Trumpf 2013-03-24 12:38:48 UTC
Following the recent commit

448efe65f51b996205dba009b0bdf25c19020d6f
soc-family: fix SOC_FAMILY override order

the contents of the variable MACHINEOVERRIDES are unfortunately still inconsistent (at least as far as I understand). It now lists SOC_FAMILY first, then MACHINE and then the active TUNE_FEATURES. 

Since the TUNE_FEATURES (armv7a in my case) are less specific than both SOC_FAMILY (omap3) and MACHINE (overo), they should come first, not last. 

This means, all the MACHINEOVERRIDES .= "${@bb.utils.contains("TUNE_FEATURES", [...] entries in meta/conf/machine/include/(ia32|arm)/*inc would need to be changed to '=.'.

Additionally, this will only work if machine config files always read soc-family.inc first and then the tune files. The overo.conf file in meta-gumstix does it in the correct order, I have not checked any other machine configs.

Not sure if there would be a better fix that does not depend on reading .inc files in a specific order. If not, this should be documented somewhere so people know to include soc-family.inc first.

The easiest way to check your install for this is using
bitbake -e | grep MACHINEOVERRIDES
Comment 1 Richard Purdie 2013-03-25 16:32:02 UTC
Patch out for review on OE-Core list.