On 2022/6/25 15:51, Li, Zhijian wrote: >> I think Li's point is still valid, but let's take action for it later. >> One more >> point I want to mention is that "unset SKIP_REASON" is not a good >> practice. To >> seek for the best shape, I can think of following changes: > > below changes sound good :) > Hi Li, Shinichiro Sorry for the late reply. Agreed. I also think "unset SKIP_REASON" is a temporary solution (not a better solution). I will try to do the following change as you suggested. Best Regards, Xiao Yang > >> >> 1) Introduce _check_kernel_option(), which checks the specified kernel >> option >> is defined, but does not set SKIP_RESAON. Using this, "unset >> SKIP_REASON" of >> group_requires() in tests/nvme-of/rc (and tests/nvme/039) can be >> avoided. >> >> 2) Introduce _have_kernel_config_file() which sets SKIP_REASON when >> neither >> /boot/config* nor /proc/config.gz is available. It can be called >> from the >> group_requires() in tests/nvme-of/rc before _check_kernel_option() >> to ensure >> the kernel option check is valid.
