Date:   Thu, 6 May 2021 20:34:48 +0800
From:   Yunsheng Lin <linyunsheng@...wei.com>
To:     Ilias Apalodimas <ilias.apalodimas@...aro.org>
CC:     Matteo Croce <mcroce@...ux.microsoft.com>,
        <netdev@...r.kernel.org>, <linux-mm@...ck.org>,
        Ayush Sawal <ayush.sawal@...lsio.com>,
        "Vinay Kumar Yadav" <vinay.yadav@...lsio.com>,
        Rohit Maheshwari <rohitm@...lsio.com>,
        "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        Thomas Petazzoni <thomas.petazzoni@...tlin.com>,
        Marcin Wojtas <mw@...ihalf.com>,
        Russell King <linux@...linux.org.uk>,
        Mirko Lindner <mlindner@...vell.com>,
        Stephen Hemminger <stephen@...workplumber.org>,
        "Tariq Toukan" <tariqt@...dia.com>,
        Jesper Dangaard Brouer <hawk@...nel.org>,
        "Alexei Starovoitov" <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        "John Fastabend" <john.fastabend@...il.com>,
        Boris Pismenny <borisp@...dia.com>,
        Arnd Bergmann <arnd@...db.de>,
        Andrew Morton <akpm@...ux-foundation.org>,
        "Peter Zijlstra (Intel)" <peterz@...radead.org>,
        Vlastimil Babka <vbabka@...e.cz>, Yu Zhao <yuzhao@...gle.com>,
        Will Deacon <will@...nel.org>,
        Michel Lespinasse <walken@...gle.com>,
        Fenghua Yu <fenghua.yu@...el.com>,
        Roman Gushchin <guro@...com>, Hugh Dickins <hughd@...gle.com>,
        Peter Xu <peterx@...hat.com>, Jason Gunthorpe <jgg@...pe.ca>,
        Guoqing Jiang <guoqing.jiang@...ud.ionos.com>,
        Jonathan Lemon <jonathan.lemon@...il.com>,
        Alexander Lobakin <alobakin@...me>,
        Cong Wang <cong.wang@...edance.com>, wenxu <wenxu@...oud.cn>,
        Kevin Hao <haokexin@...il.com>,
        Aleksandr Nogikh <nogikh@...gle.com>,
        Jakub Sitnicki <jakub@...udflare.com>,
        Marco Elver <elver@...gle.com>,
        Willem de Bruijn <willemb@...gle.com>,
        Miaohe Lin <linmiaohe@...wei.com>,
        Guillaume Nault <gnault@...hat.com>,
        <linux-kernel@...r.kernel.org>, <linux-rdma@...r.kernel.org>,
        <bpf@...r.kernel.org>, Matthew Wilcox <willy@...radead.org>,
        Eric Dumazet <edumazet@...gle.com>,
        David Ahern <dsahern@...il.com>,
        Lorenzo Bianconi <lorenzo@...nel.org>,
        Saeed Mahameed <saeedm@...dia.com>,
        Andrew Lunn <andrew@...n.ch>, Paolo Abeni <pabeni@...hat.com>
Subject: Re: [PATCH net-next v3 0/5] page_pool: recycle buffers

On 2021/5/1 0:24, Ilias Apalodimas wrote:
> [...]
>>>>
>>>> 1. skb frag page recycling do not need "struct xdp_rxq_info" or
>>>>    "struct xdp_mem_info" to bond the relation between "struct page" and
>>>>    "struct page_pool", which seems uncessary at this point if bonding
>>>>    a "struct page_pool" pointer directly in "struct page" does not cause
>>>>    space increasing.
>>>
>>> We can't do that. The reason we need those structs is that we rely on the
>>> existing XDP code, which already recycles it's buffers, to enable
>>> recycling.  Since we allocate a page per packet when using page_pool for a
>>> driver , the same ideas apply to an SKB and XDP frame. We just recycle the
>>
>> I am not really familar with XDP here, but a packet from hw is either a
>> "struct xdp_frame/xdp_buff" for XDP or a "struct sk_buff" for TCP/IP stack,
>> a packet can not be both "struct xdp_frame/xdp_buff" and "struct sk_buff" at
>> the same time, right?
>>
> 
> Yes, but the payload is irrelevant in both cases and that's what we use
> page_pool for.  You can't use this patchset unless your driver usues
> build_skb().  So in both cases you just allocate memory for the payload and

I am not sure I understood why build_skb() matters here. If the head data of
a skb is a page frag and is from page pool, then it's page->signature should be
PP_SIGNATURE, otherwise it's page->signature is zero, so a recyclable skb does
not require it's head data being from a page pool, right?

> decide what the wrap the buffer with (XDP or SKB) later.

[...]

>>
>> I am not sure I understand what you meant by "free the skb", does it mean
>> that kfree_skb() is called to free the skb.
> 
> Yes
> 
>>
>> As my understanding, if the skb completely own the page(which means page_count()
>> == 1) when kfree_skb() is called, __page_pool_put_page() is called, otherwise
>> page_ref_dec() is called, which is exactly what page_pool_atomic_sub_if_positive()
>> try to handle it atomically.
>>
> 
> Not really, the opposite is happening here. If the pp_recycle bit is set we
> will always call page_pool_return_skb_page().  If the page signature matches
> the 'magic' set by page pool we will always call xdp_return_skb_frame() will
> end up calling __page_pool_put_page(). If the refcnt is 1 we'll try
> to recycle the page.  If it's not we'll release it from page_pool (releasing
> some internal references we keep) unmap the buffer and decrement the refcnt.

Yes, I understood the above is what the page pool do now.

But the question is who is still holding an extral reference to the page when
kfree_skb()? Perhaps a cloned and pskb_expand_head()'ed skb is holding an extral
reference to the same page? So why not just do a page_ref_dec() if the orginal skb
is freed first, and call __page_pool_put_page() when the cloned skb is freed later?
So that we can always reuse the recyclable page from a recyclable skb. This may
make the page_pool_destroy() process delays longer than before, I am supposed the
page_pool_destroy() delaying for cloned skb case does not really matters here.

If the above works, I think the samiliar handling can be added to RX zerocopy if
the RX zerocopy also hold extral references to the recyclable page from a recyclable
skb too?

> 
> [1] https://lore.kernel.org/netdev/154413868810.21735.572808840657728172.stgit@firesoul/
> 
> Cheers
> /Ilias
> 
> .
> 

Powered by blists - more mailing lists