Re: [PATCH] mm,fs: Split dump_mapping() out from dump_page()

[Matthew Wilcox <willy@infradead.org>]

On Mon, Nov 22, 2021 at 08:11:32AM +0100, Michal Hocko wrote:
> On Sun 21-11-21 12:10:56, Matthew Wilcox wrote:
> > dump_mapping() is a big chunk of dump_page(), and it'd be handy to be
> > able to call it when we don't have a struct page.  Split it out and move
> > it to fs/inode.c.  Take the opportunity to simplify some of the debug
> > messages a little.
> 
> Makes sense. I haven't checked the head files inclusion side of this but
> I suspect mm heads do include uaccess.h. Not sure inode.c does as well.

It does ... convolutedly:

linux/fs.h
linux/percpu-rwsem.h
linux/rcuwait.h
linux/sched/signal.h
linux/sched/task.h
linux/uaccess.h

There may be shorter paths to including that file, but that's the
one that's actually used according to cpp.  fs/inode.c includes linux/mm.h,
so it also gets it through

linux/pgtable.h
x86/include/asm/pgtable.h
x86/include/asm/pkru.h
x86/include/asm/fpu/xstate.h

which is probably not how you think the mm files get uaccess.h, but
again, that's the first way that cpp tells me it gets pulled in.  Our
header files remain a mess.

> > Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
> 
> Acked-by: Michal Hocko <mhocko@suse.com>

Thanks!