Hi all, I see that the image_types.bbclass's IMAGE_CMD_ubi() does the following: 1) Creates a single-volume ubinize.cfg on the fly 2) Runs mkfs.ubifs to create the UBIFS image 3) Run ubinize to create a UBI image
However, I'm currently using a two-volume image: one volume for the rootfs, and one for "user config" as shown in the below ubinize.cfg. I'm considering making some changes to image_types.bbclass to support this and was hoping I could get some feedback and advice. [rootfs] mode=ubi image=rootfs.ubifs vol_id=0 vol_size=200MiB vol_type=dynamic vol_flags=autoresize vol_name=rootfs [config] mode=ubi image=config.ubifs vol_id=1 vol_size=8MiB vol_type=dynamic vol_name=config I'd planned to stage my config.ubifs with some sane default files via a recipe (e.g., an ubinize invocation in a do_compile). It feels like it might be "bad practice" to have a recipe install config.ubifs (perhaps a name with a timestamp and then a shorthand symlink, as done with existing image files) into $(DEPLOY_DIR_IMAGE}. Any recommendations? Where would be an appropriate place to keep a custom ubinize.cfg? In my case, it should always remain the same for my MACHINE. However, I'm not certain if this would be always be the case for other folks... Thank you, Jon _______________________________________________ Openembedded-core mailing list [email protected] http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
