Bitbake cudatoolkit6.5 for TK1


I am testing the GPU access on my TK1 board. My test programs are running but i am not able to integrate cudatoolkit6.5 into my image.
Bitbak’ing cudatoolkit6.5 fails with the output below. I am pretty sure that the recipe worked with BSP2.6.1. Have i done something wrong? Is there a known issue about this recipe broken?
Thanks in advance!

ERROR: cudatoolkit6.5-21.4.0-r0 do_unpack: Error executing a python function in exec_python_func() autogenerated:

The stack trace of python calls that resulted in this exception/failure was:
File: 'exec_python_func() autogenerated', lineno: 2, function: <module>
 *** 0002:do_unpack(d)
File: '/home/user/oe-core/build/../layers/meta-jetson-tk1/recipes-bsp/binary-drivers/', lineno: 33, function: do_unpack
     0030:INSANE_SKIP_${PN} += "dev-so"
     0032:python do_unpack () {
 *** 0033:    locals = d.getVar('WORKDIR')
     0034:    d.setVar('WORKDIR', '${WORKDIR}/cuda6.5')
     0035:    d.setVar('S', '${WORKDIR}')
     0036:'base_do_unpack', d)
     0037:    d.setVar('WORKDIR', locals)
Exception: TypeError: getVar() missing 1 required positional argument: 'expand'

ERROR: cudatoolkit6.5-21.4.0-r0 do_unpack: Function failed: do_unpack
ERROR: Logfile of failure stored in: /home/user/oe-core/build/tmp-glibc/work/apalis_tk1-angstrom-linux-gnueabi/cudatoolkit6.5/21.4.0-r0/temp/log.do_unpack.26434
ERROR: Task (/home/user/oe-core/build/../layers/meta-jetson-tk1/recipes-bsp/binary-drivers/ failed with exit code '1'


Have i done something wrong?

The troubleshooting section could help you finding answers to this kind of issues.

Is there a known issue about this recipe broken?

The issue is fixed in meta-toradex-tegra, morty-next branch.


Thanks! The recipes works again.

Please note that those recipes still pull the R21.4 version while the R21.5 would be available from NVIDIA. We plan to update this for our upcoming Q1 release.