From: Umang Pokhriyal <umangpokhriyall@gmail.com>
To: netdev-bot+sashiko@kernel.org
Cc: netdev@vger.kernel.org,
Willem de Bruijn <willemdebruijn.kernel@gmail.com>,
Jason Wang <jasowangio@gmail.com>,
Andrew Lunn <andrew+netdev@lunn.ch>,
"David S. Miller" <davem@davemloft.net>,
Eric Dumazet <edumazet@kernel.org>,
Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH net] tun: delete the flow entries of the moved queue on detach
Date: Sat, 14 Nov 2026 18:51:50 +0530 [thread overview]
Message-ID: <20261114132237.70664-1-umangpokhriyall@gmail.com> (raw)
In-Reply-To: <179415734502.994134.10139593761656069984@kernel.org>
On Sun, 08 Nov 2026 17:02:25 +0000, netdev-bot+sashiko@kernel.org wrote:
> Can a concurrent write() still put back an entry that points at the old
> index after this delete has run?
Yes. On the write() path tun_flow_update() reads tfile->queue_index
outside any RCU section, and synchronize_net() does not wait for it.
The race predates this patch. Before the patch the entries survived
every detach. With it, only a write() that stalls across the
synchronize_net() can store the old index again.
> There are two possible fixes. One is to read tfile->queue_index and update
> e->queue_index under tun->lock. The other is to have
> tun_automq_select_queue() fall back to reciprocal_scale() when
> e->queue_index >= numqueues. Would either of these work here?
The second one covers both sequences and keeps the write() path
lockless for a known flow. It would be a separate patch for net-next,
since the race is old and has not been seen in practice.
next prev parent reply other threads:[~2026-11-14 13:24 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-04 17:42 Umang Pokhriyal
2026-11-06 16:27 ` Willem de Bruijn
2026-11-08 17:02 ` netdev-bot+sashiko
2026-11-14 13:21 ` Umang Pokhriyal [this message]
2026-11-10 20:40 ` patchwork-bot+netdevbpf
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=20261114132237.70664-1-umangpokhriyall@gmail.com \
--to=umangpokhriyall@gmail.com \
--cc=andrew+netdev@lunn.ch \
--cc=davem@davemloft.net \
--cc=edumazet@kernel.org \
--cc=jasowangio@gmail.com \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev-bot+sashiko@kernel.org \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=willemdebruijn.kernel@gmail.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
all inboxes | Powered by JetHome®