From: Alexander H Duyck <alexander.duyck@gmail.com>
To: Felix Fietkau <nbd@nbd.name>,
Ilias Apalodimas <ilias.apalodimas@linaro.org>
Cc: netdev@vger.kernel.org, Jesper Dangaard Brouer <hawk@kernel.org>,
"David S. Miller" <davem@davemloft.net>,
Eric Dumazet <edumazet@google.com>,
Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
Lorenzo Bianconi <lorenzo@kernel.org>,
linux-kernel@vger.kernel.org,
Yunsheng Lin <linyunsheng@huawei.com>
Subject: Re: [PATCH] net: page_pool: fix refcounting issues with fragmented allocation
Date: Tue, 24 Jan 2023 13:10:24 -0800 [thread overview]
Message-ID: <cd35316065cfe8d706ca2730babe3e6519df6034.camel@gmail.com> (raw)
In-Reply-To: <19121deb-368f-9786-8700-f1c45d227a4c@nbd.name>
On Tue, 2023-01-24 at 18:22 +0100, Felix Fietkau wrote:
> On 24.01.23 15:11, Ilias Apalodimas wrote:
> > Hi Felix,
> >
> > ++cc Alexander and Yunsheng.
> >
> > Thanks for the report
> >
> > On Tue, 24 Jan 2023 at 14:43, Felix Fietkau <nbd@nbd.name> wrote:
> > >
> > > While testing fragmented page_pool allocation in the mt76 driver, I was able
> > > to reliably trigger page refcount underflow issues, which did not occur with
> > > full-page page_pool allocation.
> > > It appears to me, that handling refcounting in two separate counters
> > > (page->pp_frag_count and page refcount) is racy when page refcount gets
> > > incremented by code dealing with skb fragments directly, and
> > > page_pool_return_skb_page is called multiple times for the same fragment.
> > >
> > > Dropping page->pp_frag_count and relying entirely on the page refcount makes
> > > these underflow issues and crashes go away.
> > >
> >
> > This has been discussed here [1]. TL;DR changing this to page
> > refcount might blow up in other colorful ways. Can we look closer and
> > figure out why the underflow happens?
> I don't see how the approch taken in my patch would blow up. From what I
> can tell, it should be fairly close to how refcount is handled in
> page_frag_alloc. The main improvement it adds is to prevent it from
> blowing up if pool-allocated fragments get shared across multiple skbs
> with corresponding get_page and page_pool_return_skb_page calls.
>
> - Felix
>
Do you have the patch available to review as an RFC? From what I am
seeing it looks like you are underrunning on the pp_frag_count itself.
I would suspect the issue to be something like starting with a bad
count in terms of the total number of references, or deducing the wrong
amount when you finally free the page assuming you are tracking your
frag count using a non-atomic value in the driver.
Thanks,
- Alex
next prev parent reply other threads:[~2023-01-24 21:10 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-01-24 12:43 [PATCH] net: page_pool: fix refcounting issues with fragmented allocation Felix Fietkau
2023-01-24 14:11 ` Ilias Apalodimas
2023-01-24 15:57 ` Alexander H Duyck
2023-01-24 16:59 ` Felix Fietkau
2023-01-26 10:31 ` Ilias Apalodimas
2023-01-26 15:41 ` Alexander Duyck
2023-01-26 16:05 ` Ilias Apalodimas
2023-01-24 17:22 ` Felix Fietkau
2023-01-24 21:10 ` Alexander H Duyck [this message]
2023-01-24 21:30 ` Felix Fietkau
2023-01-25 17:11 ` Alexander H Duyck
2023-01-25 17:32 ` Felix Fietkau
2023-01-25 18:26 ` Alexander H Duyck
2023-01-25 18:42 ` Felix Fietkau
2023-01-25 19:02 ` Alexander H Duyck
2023-01-25 19:10 ` Felix Fietkau
2023-01-25 19:40 ` Felix Fietkau
2023-01-25 20:02 ` Felix Fietkau
2023-01-25 22:14 ` Alexander H Duyck
2023-01-26 6:12 ` Felix Fietkau
2023-01-26 9:14 ` Felix Fietkau
2023-01-26 16:08 ` Alexander Duyck
2023-01-26 16:40 ` Alexander Duyck
2023-01-26 17:44 ` Felix Fietkau
2023-01-26 18:38 ` Alexander H Duyck
2023-01-26 18:43 ` Felix Fietkau
2023-01-26 19:06 ` [net PATCH] skb: Do mix page pool and page referenced frags in GRO Alexander Duyck
2023-01-26 19:14 ` Toke Høiland-Jørgensen
2023-01-26 19:48 ` Alexander Duyck
2023-01-26 21:35 ` Toke Høiland-Jørgensen
2023-01-26 23:13 ` Jakub Kicinski
2023-01-27 7:15 ` Ilias Apalodimas
2023-01-27 7:21 ` Felix Fietkau
2023-01-30 16:49 ` Jesper Dangaard Brouer
2023-01-28 2:37 ` Yunsheng Lin
2023-01-28 5:26 ` Jakub Kicinski
2023-01-28 7:08 ` Eric Dumazet
2023-01-30 8:50 ` Paolo Abeni
2023-01-30 16:17 ` Alexander Duyck
2023-01-28 7:15 ` Eric Dumazet
2023-01-28 17:08 ` Alexander Duyck
2023-01-28 7:50 ` patchwork-bot+netdevbpf
2023-01-26 10:32 ` [PATCH] net: page_pool: fix refcounting issues with fragmented allocation 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=cd35316065cfe8d706ca2730babe3e6519df6034.camel@gmail.com \
--to=alexander.duyck@gmail.com \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=hawk@kernel.org \
--cc=ilias.apalodimas@linaro.org \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linyunsheng@huawei.com \
--cc=lorenzo@kernel.org \
--cc=nbd@nbd.name \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
/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).