Hi Nicholas,
I was looking at all the previous discussions at [1] and [2], wherein
vmalloc-huge was
switched on by default, and then had to be switched off on all arches
after an year...
a) I am wondering whether ppc supports vmalloc-huge by default right now?
It doesn't look like it since the VM_ALLOW_HUGE_VMAP flag is still there.
b) From reading all the discussions, I could only figure two problems
with vmalloc-huge - one
being the drivers operating on tail pages of a large page, thus needing
us to split the compound page,
which has been solved. The other being partial permission change. Is
there any other case I am missing,
or any other problems you speculate?
[1] https://lore.kernel.org/all/20220422060107.781512-2-npig...@gmail.com/
[2] https://lore.kernel.org/all/20220415164413.2727220-1-s...@kernel.org/
