If you set autoconf up to allow PARALLEL_MAKE ?= "-j ${@oe.utils.cpu_count()}", then odd things happen. Specifically: NOTE: recipe quilt-native-0.64-r0: task do_compile: Started ERROR: oe_runmake failed ERROR: Function failed: do_compile (log file is located at /home/pidge/yocto-autobuilder/yocto-worker/nightly/build/build-brcm97449svms-refboard/tmp/work/x86_64-linux/quilt-native/0.64-r0/temp/log.do_compile.16490) ERROR: Logfile of failure stored in: /home/pidge/yocto-autobuilder/yocto-worker/nightly/build/build-brcm97449svms-refboard/tmp/work/x86_64-linux/quilt-native/0.64-r0/temp/log.do_compile.16490 Log data follows: | DEBUG: Executing shell function do_compile | NOTE: make 16 | make: *** No rule to make target '16'. Stop. | ERROR: oe_runmake failed | ERROR: Function failed: do_compile (log file is located at /home/pidge/yocto-autobuilder/yocto-worker/nightly/build/build-brcm97449svms-refboard/tmp/work/x86_64-linux/quilt-native/0.64-r0/temp/log.do_compile.16490) NOTE: recipe quilt-native-0.64-r0: task do_compile: Failed It seems the -j is getting stripped out somewhere. If you set the PARALLEL_MAKE to something like "-j 4" then everything is fine. This might be happening in bitbake, as this works fine in shell (obviously).
If you are putting this in config/autobuilder.conf, the value for PARALLEL_MAKE is just the number with no "-j". The "-j" is added later in CreateAutoConf.py. This confused me no end at one point and I got that same exact 'make 16' error.
Beth: Was this a configuration error as per Bill's question?
Well, so let me explain my setup. I do not run CreateAutoConf. It's taken care of by some external processes. I run a somewhat hacked RunPreamble that does the setup, and then I run BuildImages. So, yeah, I don't think it's coming from autobuilder.conf. I guess the way to test this is set an append in CreateAutoConf: PARALLEL_MAKE = "-j ${@oe.utils.cpu_count()}"
New autobuilder code doesn't have this issue.