Re: [mm, net-next v2] mm: net: memcg accounting for TCP rx zerocopy

[Arjun Roy]

On Mon, Mar 15, 2021 at 11:22 PM Arjun Roy <arjunroy@xxxxxxxxxx> wrote:
>
> On Mon, Mar 15, 2021 at 9:29 PM Shakeel Butt <shakeelb@xxxxxxxxxx> wrote:
> >
> > On Mon, Mar 15, 2021 at 9:20 PM Arjun Roy <arjunroy@xxxxxxxxxx> wrote:
> > >
> > [...]
> > > >
> > >
> > > Apologies for the spam - looks like I forgot to rebase the first time
> > > I sent this out.
> > >
> > > Actually, on a related note, it's not 100% clear to me whether this
> > > patch (which in its current form, applies to net-next) should instead
> > > be based on the mm branch - but the most recent (clean) checkout of mm
> > > fails to build for me so net-next for now.
> > >
> >
> > It is due to "mm: page-writeback: simplify memcg handling in
> > test_clear_page_writeback()" patch in the mm tree. You would need to
> > reintroduce the lock_page_memcg() which returns the memcg and make
> > __unlock_page_memcg() non-static.
>
> To clarify, Shakeel - the tag "tag: v5.12-rc2-mmots-2021-03-11-21-49"
> fails to build on a clean checkout, without this patch, due to a
> compilation failure in mm/shmem.c, for reference:
> https://pastebin.com/raw/12eSGdGD
> (and that's why I'm basing this patch off of net-next in this email).
>
> -Arjun

Another seeming anomaly - the patch sent out passes
scripts/checkpatch.pl but netdev/checkpatch finds plenty of actionable
fixes here: https://patchwork.kernel.org/project/netdevbpf/patch/20210316041645.144249-1-arjunroy.kdev@xxxxxxxxx/

Is netdev using some other automated checker instead of scripts/checkpatch.pl?

-Arjun