On Fri, Oct 16, 2020 at 11:19 AM Vivek Goyal <vgoyal@xxxxxxxxxx> wrote:
>
> So with multi page writes performance seems much better for this
> particular workload.
Looking at that code, I don't see why the page needs to be unlocked
after it's been filled, so I do think the easy solution is to just
always unlock the pages. Then the rest of the multi-page code is
probably fine.
But hey, who knows. I might be missing something.
Linus
