On Mon, Aug 26, 2024 at 04:44:01PM GMT, Kent Overstreet wrote:

No, I explained why GFP_NORECLAIM/PF_MEMALLOC_NORECLAIM can absolutely
apply to a context, not a callsite, and why vmalloc() and kvmalloc()
ignoring gfp flags is a much more serious issue.

If you want to do something useful, figure out what we're going to do
about _that_. If you really don't want PF_MEMALLOC_NORECLAIM to exist,
then see if Linus will let you plumb gfp flags down to pte allocation -
and beware, that's arch code that you'll have to fix.

Reminder: kvmalloc() is a thing, and it's steadily seeing wider use.

Otherwise, PF_MEMALLOC_NORECLAIM needs to stay; and thank you for
bringing this to my attention, because it's made me realize all the
other places in bcachefs that use gfp flags for allocating memory with
btree locks held need to be switch to memalloc_flags_do().

Additionally: plumbing gfp flags to pte allocation is something we do
need to do. I proposed it before kvmalloc() was a thing, but now it's
become much more of a lurking landmine.

Even with that I'd still be against this series, though. GFP_NOFAIL
interacting badly with other gfp/memalloc flags is going to continue to
be an issue, and I think the only answer to that is stricter runtime
checks (which, thank you mm guys for adding recently).