On Tue, 30 Aug 2022 16:23:44 +0200 Alexander Potapenko <glider@xxxxxxxxxx> wrote:
> > from init/do_mounts.c:2:
> > ./include/linux/page-flags.h: In function ‘page_fixed_fake_head’:
> > ./include/linux/page-flags.h:226:36: error: invalid use of undefined type ‘const struct page’
> > 226 | test_bit(PG_head, &page->flags)) {
> > | ^~
> > ./include/linux/bitops.h:50:44: note: in definition of macro ‘bitop’
> > 50 | __builtin_constant_p((uintptr_t)(addr) != (uintptr_t)NULL) && \
> > | ^~~~
> > ./include/linux/page-flags.h:226:13: note: in expansion of macro ‘test_bit’
> > 226 | test_bit(PG_head, &page->flags)) {
> > | ^~~~~~~~
> > ...
>
> Gotcha, this is a circular dependency: mm_types.h -> sched.h ->
> kmsan.h -> gfp.h -> mmzone.h -> page-flags.h -> mm_types.h, where the
> inclusion of sched.h into mm_types.h was only introduced in "mm:
> multi-gen LRU: support page table walks" - that's why the problem was
> missing in other trees.
Ah, thanks for digging that out.
Yu, that inclusion is regrettable. I don't think mm_types.h is an
appropriate site for implementing lru_gen_use_mm() anyway. Adding a
new header is always the right fix for these things. I'd suggest
adding a new mglru.h (or whatever) and putting most/all of the mglru
material in there.
Also, the addition to kernel/sched/core.c wasn't clearly changelogged,
is uncommented and I doubt if the sched developers know about it, let
alone reviewed it. Please give them a heads-up.
The addition looks fairly benign, but core context_switch() is the
sort of thing which people get rather defensive about and putting
mm-specific stuff in there might be challenged. Some quantitative
justification of this optimization would be appropriate.
> In fact sched.h only needs the definitions of `struct
> kmsan_context_state` and `struct kmsan_ctx` from kmsan.h, so I am
> splitting them off into kmsan_types.h to break this circle.
> Doing so also helped catch a couple of missing/incorrect inclusions of
> KMSAN headers in subsystems.
