Date:   Wed, 24 Mar 2021 13:53:34 -0700
From:   Shakeel Butt <shakeelb@...gle.com>
To:     Arjun Roy <arjunroy@...gle.com>
Cc:     Michal Hocko <mhocko@...e.com>,
        Johannes Weiner <hannes@...xchg.org>,
        Arjun Roy <arjunroy.kdev@...il.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        David Miller <davem@...emloft.net>,
        netdev <netdev@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Cgroups <cgroups@...r.kernel.org>, Linux MM <linux-mm@...ck.org>,
        Eric Dumazet <edumazet@...gle.com>,
        Soheil Hassas Yeganeh <soheil@...gle.com>,
        Jakub Kicinski <kuba@...nel.org>,
        Yang Shi <shy828301@...il.com>, Roman Gushchin <guro@...com>
Subject: Re: [mm, net-next v2] mm: net: memcg accounting for TCP rx zerocopy

On Wed, Mar 24, 2021 at 1:39 PM Arjun Roy <arjunroy@...gle.com> wrote:
>
> On Wed, Mar 24, 2021 at 2:12 AM Michal Hocko <mhocko@...e.com> wrote:
> >
> > On Tue 23-03-21 11:47:54, Arjun Roy wrote:
> > > On Tue, Mar 23, 2021 at 7:34 AM Michal Hocko <mhocko@...e.com> wrote:
> > > >
> > > > On Wed 17-03-21 18:12:55, Johannes Weiner wrote:
> > > > [...]
> > > > > Here is an idea of how it could work:
> > > > >
> > > > > struct page already has
> > > > >
> > > > >                 struct {        /* page_pool used by netstack */
> > > > >                         /**
> > > > >                          * @dma_addr: might require a 64-bit value even on
> > > > >                          * 32-bit architectures.
> > > > >                          */
> > > > >                         dma_addr_t dma_addr;
> > > > >                 };
> > > > >
> > > > > and as you can see from its union neighbors, there is quite a bit more
> > > > > room to store private data necessary for the page pool.
> > > > >
> > > > > When a page's refcount hits zero and it's a networking page, we can
> > > > > feed it back to the page pool instead of the page allocator.
> > > > >
> > > > > From a first look, we should be able to use the PG_owner_priv_1 page
> > > > > flag for network pages (see how this flag is overloaded, we can add a
> > > > > PG_network alias). With this, we can identify the page in __put_page()
> > > > > and __release_page(). These functions are already aware of different
> > > > > types of pages and do their respective cleanup handling. We can
> > > > > similarly make network a first-class citizen and hand pages back to
> > > > > the network allocator from in there.
> > > >
> > > > For compound pages we have a concept of destructors. Maybe we can extend
> > > > that for order-0 pages as well. The struct page is heavily packed and
> > > > compound_dtor shares the storage without other metadata
> > > >                                         int    pages;    /*    16     4 */
> > > >                         unsigned char compound_dtor;     /*    16     1 */
> > > >                         atomic_t   hpage_pinned_refcount; /*    16     4 */
> > > >                         pgtable_t  pmd_huge_pte;         /*    16     8 */
> > > >                         void *     zone_device_data;     /*    16     8 */
> > > >
> > > > But none of those should really require to be valid when a page is freed
> > > > unless I am missing something. It would really require to check their
> > > > users whether they can leave the state behind. But if we can establish a
> > > > contract that compound_dtor can be always valid when a page is freed
> > > > this would be really a nice and useful abstraction because you wouldn't
> > > > have to care about the specific type of page.
> > > >
> > > > But maybe I am just overlooking the real complexity there.
> > > > --
> > >
> > > For now probably the easiest way is to have network pages be first
> > > class with a specific flag as previously discussed and have concrete
> > > handling for it, rather than trying to establish the contract across
> > > page types.
> >
> > If you are going to claim a page flag then it would be much better to
> > have it more generic. Flags are really scarce and if all you care about
> > is PageHasDestructor() and provide one via page->dtor then the similar
> > mechanism can be reused by somebody else. Or does anything prevent that?
>
> The way I see it - the fundamental want here is, for some arbitrary
> page that we are dropping a reference on, to be able to tell that the
> provenance of the page is some network driver's page pool. If we added
> an enum target to compound_dtor, if we examine that offset in the page
> and look at that value, what guarantee do we have that the page isn't
> instead some other kind of page, and the byte value there was just
> coincidentally the one we were looking for (but it wasn't a network
> driver pool page)?
>
> Existing users of compound_dtor seem to check first that a
> PageCompound() or PageHead() return true - the specific scenario here,
> of receiving network packets, those pages will tend to not be compound
> (and more specifically, compound pages are explicitly disallowed for
> TCP receive zerocopy).
>
> Given that's the case, the options seem to be:
> 1) Use a page flag - with the downside that they are a severely
> limited resource,
> 2) Use some bits inside page->memcg_data - this I believe Johannes had
> reasons against, and it isn't always the case that MEMCG support is
> enabled.
> 3) Use compound_dtor - but I think this would have problems for the
> prior reasons.

I don't think Michal is suggesting to use PageCompound() or
PageHead(). He is suggesting to add a more general page flag
(PageHasDestructor) and corresponding page->dtor, so other potential
users can use it too.

Powered by blists - more mailing lists