* Borislav Petkov wrote: > thanks for the report. > > So, AFAICT, this is caused by ksysfs ioremapping struct > setup_data for a short time so that it can count it and > show it in /sys/kernel/boot_params/setup_data/* > > And, of course, the setup_data thing which we're using > for kaslr param passing is RAM and ioremap complains. > > And currently I don't have a good idea how to fix it. > Perhaps introduce an ioremap_* something which suppresses > the warning as we're going to iounmap() right afterwards > but that's ugly. Why is it ioremap()-ed to begin with, why cannot the kernel access its own data structure in RAM directly? Thanks, Ingo