IMO things we want to have in future should be #ifdef-ed in driver rather than dummy-implemented in LKPI
But in that case we can define opaque struct device_node; in linux/of.h to satisfy struct device_node *of_node; member of struct class and than add dummy implementation somewhere in wifi driver to get rid of extra #ifdef-s

lkpi->linuxkpi for public interfaces
move function declarations to beginning of pci.h and sort them in order of definition
remove forgotten inlines
make IRQF_SHARED flag FreeBSD-agnostic

allow end parameter of ida_simple_get() be larger than INT_MAX. Linux caps the value to INT_MAX.

remove duplicated typeof_member from linux/kernel.h

Allow new functions to work with mutex-less seqcounts

Fix typo

In D42821#979762, @bz wrote:

How does that lock prevent us from any other changes at the same time by anything calling pcie_capability_write_word() in between?

Add locking in pcie_capability_set_word() and pcie_capability_clear_word()

Fix typo

Use drm-kmod version of linux/ioport.h placed in https://github.com/freebsd/drm-kmod/blob/master/linuxkpi/gplv2/include/linux/pci.h as a base.
It properly defines members of struct resource.

READ_ONCE -> atomic_load_int

INT_MAX -> UINT_MAX

Make IOMEM_ERR_PTR sparse-friendly and convert to macro to match Linux

Change copyright owners to match git blame output

Add LINUXKPI to guard #define
style(9)
typeof -> __typeof

Add prequisite includes

Add LINUXKPI to guard #define
Use FreeBSD container_of implementation
typeof -> __typeof
Not based on OpenBSD version now

Add LINUXKPI to guard #define

Squashed in to D42793

Add LINUXKPI to guard #define
Squash with D42813

Now list_add() calls __list_add() by it Linux name

rename old linux_list_add function instead of adding a new alias

The aim of this set of reviews is to reach 6.1-lts (6.1.63 at the moment of writing). https://github.com/wulf7/drm-kmod/commits/6.1-lts