On Fri, Oct 14, 2022 at 10:11 PM Jose Quaresma <quaresma.jose@...> wrote:

Hi Rudolf,

I have reported this issue and together with a path [1] to fix it.

Anyway Bruce has fixed the issue in another way in [2] and this works for me as well.

[1] https://lists.yoctoproject.org/g/linux-yocto/message/11746

[2] https://git.yoctoproject.org/yocto-kernel-tools/commit/?id=6a4752ebbe7d242c02b3c74a5772926edd243626

Yes indeed. This has also passed my other regression testing, but I was traveling today and couldn't send my pull request.

I'll send it over the weekend, so hopefully it'll be merged shortly!

Bruce

 

Jose

Rudolf J Streif <rudolf.streif@...> escreveu no dia sábado, 15/10/2022 à(s) 01:45:

I am running into a bizarre problem with kernel configuration.

I have an existing build environment that builds for an i.MX6ul machine.
That works fine and the linux-fslc kernel 5.18 builds without any
problems. Now within the that build environment I changed the machine to
imx6ullevk and then the kernel configuration fails with this error message:

| DEBUG: Executing shell function do_kernel_configme
| NOTE: do_kernel_metadata: for summary/debug, set KCONF_AUDIT_LEVEL > 0
| [ERROR]: processing of file /tmp/tmp.owmBptYrba failed
| /develop/projects/altec/tcu/build/tmp/hosttools/dirname: missing operand
| Try '/develop/projects/altec/tcu/build/tmp/hosttools/dirname --help'
for more information.
| ERROR: Could not generate configuration queue for imx6ullevk.
| WARNING:
/develop/projects/altec/tcu/build/tmp/work/imx6ullevk-altec-linux-gnueabi/linux-fslc/5.18.5+gitAUTOINC+1d6b3055ae-r0/temp/run.do_kernel_configme.287395:209
exit 1 from 'exit 1'
| WARNING: Backtrace (BB generated script):
|     #1: bbfatal_log,
/develop/projects/altec/tcu/build/tmp/work/imx6ullevk-altec-linux-gnueabi/linux-fslc/5.18.5+gitAUTOINC+1d6b3055ae-r0/temp/run.do_kernel_configme.287395,
line 209
|     #2: do_kernel_metadata,
/develop/projects/altec/tcu/build/tmp/work/imx6ullevk-altec-linux-gnueabi/linux-fslc/5.18.5+gitAUTOINC+1d6b3055ae-r0/temp/run.do_kernel_configme.287395,
line 382
|     #3: do_kernel_configme,
/develop/projects/altec/tcu/build/tmp/work/imx6ullevk-altec-linux-gnueabi/linux-fslc/5.18.5+gitAUTOINC+1d6b3055ae-r0/temp/run.do_kernel_configme.287395,
line 150
|     #4: main,
/develop/projects/altec/tcu/build/tmp/work/imx6ullevk-altec-linux-gnueabi/linux-fslc/5.18.5+gitAUTOINC+1d6b3055ae-r0/temp/run.do_kernel_configme.287395,
line 484
ERROR: Task
(/develop/projects/altec/tcu/build/../meta-freescale/recipes-kernel/linux/linux-fslc_5.18.bb:do_kernel_configme)
failed with exit code '1'

I tried to debug do_kernel_configme. The issue seems in the scc script
more specifically, I think, in the kconf command that is called by scc
when parsing the temporary configuration queue file in /tmp: tmp.owmBptYrba

#
# spp v0.8
# processed: Fri Oct 14 09:49:02 PM UTC 2022
#
# This is a preprocessor output file, do not edit
#
#
prefix
/develop/projects/altec/tcu/build/../meta-freescale/recipes-kernel/linux/linux-fslc/defconfig
kconf non-hardware
/develop/projects/altec/tcu/build/../meta-freescale/recipes-kernel/linux/linux-fslc/defconfig
# run time: 0 seconds
# processed files:
# _cfg
/develop/projects/altec/tcu/build/../meta-freescale/recipes-kernel/linux/linux-fslc/defconfig

Apparently a call to dirname is missing the argument. However, I do not
understand why.

I wiped out build/tmp directory entirely, ran a cleanall on
virtual/kernel but to no avail. The problem remains the same.

Has anybody seen this issue before?


Rudi

--

Best regards,

José Quaresma

--

- Thou shalt not follow the NULL pointer, for chaos and madness await thee at its end
- "Use the force Harry" - Gandalf, Star Trek II