On Wed, Mar 1, 2023 at 9:10 AM Matthew Wilcox (Oracle)
<willy@xxxxxxxxxxxxx> wrote:
>
> + memcpy_to_file_folio(folio, 0, vip->vii_immed.vi_immed, isize);
Well, this function doesn't exist upstream yet, much less in any
stable kernels..
In fact, I can't find any sign of that function *anywhere*. Searching
for it on lkml finds zero hits, as does linux-next.
Is it something hidden in your personal tree, or is my grep failing
because the function generation is behind some macro magic?
And while I'm on this subject: the "memcpy_from_file_folio()"
interface is horrendously broken. For the highmem case, it shouldn't
be an inline function, and it should loop over pages - instead of
leaving the callers having to do that.
Of course, callers don't actually do that (since there are no callers
- unless I'm again missing it due to some macro games with token
concatenation), but I really wish it was fixed before any callers
appear.
Linus
