From: Tariq Toukan <tariqt@mellanox.com>
To: Eric Dumazet <eric.dumazet@gmail.com>,
Ilias Apalodimas <ilias.apalodimas@linaro.org>,
Matthew Wilcox <willy@infradead.org>,
"brouer@redhat.com" <brouer@redhat.com>
Cc: David Miller <davem@davemloft.net>,
"toke@redhat.com" <toke@redhat.com>,
"netdev@vger.kernel.org" <netdev@vger.kernel.org>,
"mgorman@techsingularity.net" <mgorman@techsingularity.net>,
"linux-mm@kvack.org" <linux-mm@kvack.org>
Subject: Re: [RFC, PATCH] net: page_pool: Don't use page->private to store dma_addr_t
Date: Tue, 12 Feb 2019 12:39:59 +0000 [thread overview]
Message-ID: <27e97aac-f25b-d46c-3e70-7d0d44f784b5@mellanox.com> (raw)
In-Reply-To: <64f7af75-e6df-7abc-c4ce-82e6ca51fafe@gmail.com>
On 2/11/2019 7:14 PM, Eric Dumazet wrote:
>
>
> On 02/11/2019 12:53 AM, Tariq Toukan wrote:
>>
>
>> Hi,
>>
>> It's great to use the struct page to store its dma mapping, but I am
>> worried about extensibility.
>> page_pool is evolving, and it would need several more per-page fields.
>> One of them would be pageref_bias, a planned optimization to reduce the
>> number of the costly atomic pageref operations (and replace existing
>> code in several drivers).
>>
>
> But the point about pageref_bias is to place it in a different cache line than "struct page"
>
> The major cost is having a cache line bouncing between producer and consumer.
>
pageref_bias is meant to be dirtied only by the page requester, i.e. the
NIC driver / page_pool.
All other components (basically, SKB release flow / put_page) should
continue working with the atomic page_refcnt, and not dirty the
pageref_bias.
However, what bothers me more is another issue.
The optimization doesn't cleanly combine with the new page_pool
direction for maintaining a queue for "available" pages, as the put_page
flow would need to read pageref_bias, asynchronously, and act accordingly.
The suggested hook in put_page (to catch the 2 -> 1 "biased refcnt"
transition) causes a problem to the traditional pageref_bias idea, as it
implies a new point in which the pageref_bias field is read
*asynchronously*. This would risk missing the this critical 2 -> 1
transition! Unless pageref_bias is atomic...
> pageref_bias means the producer only have to read the "struct page" and not dirty it
> in the case the page can be recycled.
>
>
>
>> I would replace this dma field with a pointer to an extensible struct,
>> that would contain the dma mapping (and other stuff in the near future).
>> This pointer fits perfectly with the existing unsigned long private;
>> they can share the memory, for both 32- and 64-bits systems.
>>
>> The only downside is one more pointer de-reference. This should be perf
>> tested.
>> However, when introducing the page refcnt bias optimization into
>> page_pool, I believe the perf gain would be guaranteed.
>
> Only in some cases perhaps (when the cache line can be dirtied without performance hit)
>
next prev parent reply other threads:[~2019-02-12 12:40 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-02-07 14:36 [RFC, PATCH] net: page_pool: Don't use page->private to store dma_addr_t Ilias Apalodimas
2019-02-07 15:07 ` Matthew Wilcox
2019-02-07 15:20 ` Ilias Apalodimas
2019-02-07 21:25 ` David Miller
2019-02-07 21:34 ` Matthew Wilcox
2019-02-07 21:42 ` Ilias Apalodimas
2019-02-11 8:53 ` Tariq Toukan
2019-02-11 12:12 ` Matthew Wilcox
2019-02-11 15:38 ` Tariq Toukan
2019-02-11 17:14 ` Eric Dumazet
2019-02-12 12:39 ` Tariq Toukan [this message]
2019-02-12 13:49 ` Jesper Dangaard Brouer
2019-02-12 14:58 ` Tariq Toukan
2019-02-12 15:15 ` Eric Dumazet
2019-02-12 18:13 ` Alexander Duyck
2019-02-12 18:20 ` Ilias Apalodimas
2019-02-13 8:50 ` Tariq Toukan
2019-02-13 8:46 ` Tariq Toukan
2019-02-07 21:37 ` Ilias Apalodimas
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=27e97aac-f25b-d46c-3e70-7d0d44f784b5@mellanox.com \
--to=tariqt@mellanox.com \
--cc=brouer@redhat.com \
--cc=davem@davemloft.net \
--cc=eric.dumazet@gmail.com \
--cc=ilias.apalodimas@linaro.org \
--cc=linux-mm@kvack.org \
--cc=mgorman@techsingularity.net \
--cc=netdev@vger.kernel.org \
--cc=toke@redhat.com \
--cc=willy@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).