linuxkpi: gracefully handle page lookup failure in lkpi_vmf_insert_pfn_prot_locked

[ashafer]

Summary

Currently lkpi_vmf_insert_pfn_prot_locked will check the page iter to
find a usage of the page. If no page was found, it continues on to
try using PHYS_TO_VM_PAGE() to get a page. Currently it does not check
if a valid page was found before passing it to vm_page_busy_acquire,
which can cause a kernel page fault as vm_page_busy_acquire expects
a valid page pointer.

This can easily be triggered while starting KDE6 in wayland mode, which
many users have been reporting. With this change plasma6 starts properly
in wayland mode.

Sponsored by: NVIDIA

[ashafer]

+kbowling (mentor) to confirm I can submit this myself.

[alc]

sys/compat/linuxkpi/common/src/linux_page.c

349

I just saw the MFC of this change. Since the iterator, pages, is a local variable within this function that will go out of scope when the function returns, there is no need to reset it immediately before that return.

[ashafer]

sys/compat/linuxkpi/common/src/linux_page.c

349

Sorry I had thought this was a cleanup function that did more, I was matching the teardown before the goto statement below. Is this something worth removing then or leave as-is since it's a noop?

[markj]

sys/compat/linuxkpi/common/src/linux_page.c

349

Sorry I missed this. It would be better to remove it, just to avoid setting a confusing example.