19.03.2021 18:50, Greg Kroah-Hartman пишет:

quoted

Then initialization order won't be a problem.

I don't understand the problem here, what is wrong with the patch as-is?

The cma->stat is NULL at the time when CMA is used on ARM because
cma->stat is initialized at the subsys level. This is the problem,
apparently.

Also, watch out, if you only make the kobject dynamic, how are you going
to get backwards from the kobject to the values you want to send to
userspace in the show functions?

Still there should be a wrapper around the kobj with a back reference to
the cma entry. If you're suggesting that I should write a patch, then I
may take a look at it later on. Although, I assume that Minchan could
just correct this patch and re-push it to -next.